I can't stress enough how much of a scam Honey is for retailers. If you have not blacklisted Honey from your affiliate program, or set a separate commission at 1%, I suggest you do that now.
Ranking your videos organically before using them as ads helps ramp up creatives 2.5x faster than simply using an unlisted video. I tested this out several times — 90% of the time the ranked creatives ramped up 2.5x faster and more efficiently.
Paid media is like stretching a photo — the more you expand it, the blurrier it gets. Bringing organic traffic, new customer personas, and new audiences to your brand only helps your paid media work even harder for you.
Pro tip: When you see a brand running 80% of a similar ad format, that's because it's their bread and butter, and that format works.
When spending 7-figure monthly budgets on paid social, responding to ad comments from your brand page can reduce your acquisition cost by about 20% on average, sometimes more.
Don't look at isolated performance on marketing channels. Look at total customers divided by total spend across all channels. In most cases, your macro CPA drops while you get more customers.
Google and Facebook take about 40% of every VC dollar raised for customer acquisition costs.
You must be extremely native to the platform you're showing your ad on. Just like you'd never use a Snapchat story as a billboard, you shouldn't use a TV commercial on Facebook and expect it to do well.
When a new brand launches, the biggest misconception is that the second you turn Facebook ads 'on', your whole business plan just works. Facebook will magically find you every possible person who is going to buy your product. All of that is false. You still have to build a brand. Your Facebook ads are just a vehicle for getting your future customers.
Most brands today (>95%) find minimal success in channels like Facebook because they follow the 2013 playbook: run very basic ads to a homepage — like a digital billboard that leads to nowhere magical.
If you've already built lookalike audiences, find brands that have similar audiences and cross-share your lookalikes. When I started doing this with larger companies ($50-100M rev range) in 2017, it was a game-changer.
I promise with JUST doing audience targeting, creative optimization, and landing experience right, you'll drop your Facebook CPA by 30% minimum.
The average Facebook ads CTR is 0.89%, which is pretty terrible. A good sign of a CTR would be anywhere between 2-6%.
Performance PR: take your organic earned media and amplify it with paid spend, giving more life to something that had a great impact on your brand but didn't see as many eyeballs organically. It does what I call 'full-funnel marketing.'
When I first got into DTC, I quickly learned that if you can figure out how to be contextual to the platform you distribute or show up in, you'll always do well. It's not that Facebook or Instagram don't work — it's that people aren't adapting.
People love Facebook ads because the targeting is incredible and you can see the ROAS right there. When you look at any adjacent platform it feels inferior. But that's because you're judging it by Facebook's rules, not its own.
If you run a fast-cutting Instagram ad designed for sound-off that goes to your homepage, and expect it to work on any other channel, you'll never find success. But running an influencer video to a simple landing page with 3 CTA options might yield the return.
Pinterest is like Google search but on steroids. People use Pinterest to look for answers, and when they find the answer and click, they research and then shop. We used it with a millennial-mom-targeted brand, and it had the highest time on site and highest conversion rate of all channels.
When I was running display ads with a $1M/month Facebook budget, we saw that adding display really brought CPA down. But when we lowered Facebook spend, the effectiveness of the display ads lowered too. They work together, not alone.
With Taboola and Outbrain, what works is really lengthy landing pages with editorial content, listicles comparing products, or pointing to great earned media. Launch, understand which publishers work, narrow down, then ramp up spend.
With Snapchat, you'll likely have success with view-through conversions from a retargeting audience. A positive sign is seeing your time to purchase get lowered when adding Snapchat to the mix.
You might think your CRM-remarketing CPO is $13 according to the dashboard, but through incrementality testing you could realize it's more like $68.
CPCs to content are always cheaper than CPCs to a landing page.
When I ran ads from a publisher's Facebook page instead of the brand's page, our CPA was about 15-20% lower. Social media scrollers love stories, being entertained, and content — not digital billboards.
In Q4, when ad costs are expensive to drive to a landing page, clicks to good content are always cheaper. When Q1 rolls around and CPMs drop, you'll be even happier you set this strategy up to build audiences quickly.
The brands that were minimally impacted by iOS changes were those that always answered 'Why does this brand deserve to exist' in every ad, email, SMS, retargeting banner, and direct mailer. Quick-cut Instagram ads can't be your end game for driving new customers.
Someone should see different messaging and creative at each retargeting stage: hitting the site once, selecting a product, adding to cart, and beginning checkout. This should also match your site-abandon emails.
Kettle & Fire's 'prospecting' was heavily focused on SEO. They leveraged affiliate partners for backlinks, building SEO juice for 'bone broth.' Their paid media started with just retargeting site visitors.
Having a good understanding of ad platforms is smart because down the road when you DO hire an agency, you can call them on their BS when you see it.
'Fixing Facebook ads' has everything to do with things outside of Facebook Ads Manager. 9 out of 10 times, the brand hasn't figured out all the levers around the media buy itself — pixel tracking, creative, angles, landing pages, PR, organic channels.
List 25 reasons someone should buy the product. Out of those 25, at least 15 probably haven't been tested as new angles. The key is these reasons help better someone's life — not help you trick someone into buying.
Everything to get 'Facebook ads cheaper' is outside of the ads themselves. We didn't even touch the website UX and conversion optimization piece, which is also huge.
Running ads right out the gate after launch is a huge red flag — a narrative heavily pushed by venture capital dollars over the last 6 years. It simply doesn't work. Look at Casper today... there is no soul to the company.
With iOS 14.5, Apple made it significantly harder for tracking events. Facebook used to know exactly when to serve an ad based on everything people did on their phones. Now that's all gone, which is why building your own zero-party data bank is increasingly important.
No one is fully dedicated to your account at an agency. The problem with this setup, especially in lower-funnel media, is that you need someone who's refreshing the ad account and testing new things on a constant basis.
Pre-iOS 14.5, Facebook was so good at knowing what to put in front of someone that driving to a mediocre homepage could still work. The game changed from trusting Facebook on what to show, to 'We need to create a reason that someone would want to buy us and champion that from ad creative through the purchase experience.'
Overnight after iOS 14.5, acquisition costs tripled, and most people scaled back their ad budgets.
Ads that work today are less about pushing the brand with graphics and quick-cut videos, and more about teaching someone why they're missing out by not being a consumer.
In 2017, you could run a Facebook retargeting campaign and almost guarantee purchases at just a few dollars each. Running the same playbook today, the CPO comes to roughly the same as acquiring a brand new customer. That's nuts.
OOH ads prompted consumers to take action: 41% used a search engine, 33% searched social media, 33% visited a website, and 20% made an online purchase.
We put billboards on the sides of trucks, captured mobile device IDs of people in viewing range, and retargeted them on Instagram and TikTok. With an $85 AOV brand that normally has a $60-70 CPA, we saw a $14 retargeting CPA on top of the truck spend.
Programmatic TV can be a UGC compilation — it doesn't need to be a perfectly shot $100k piece of creative.
On Taboola, when we tested 50-100 sites, we found 2 that delivered CPAs at 1/3 of Facebook. That's all we spent on Taboola. Same with TV — Baby First TV was an underpriced channel. We ran a newsletter placement with 6,000 readers for two weeks for $500 total.
Using UTMs allows you to properly understand segments of traffic. In 6 months, you can see the LTV of customers who came in through an Instagram ad with a static studio shot vs a lifestyle shot.
Warren Buffett's parade analogy applied to DTC: 'One spectator stands on tiptoes. It works until everyone does the same. Then standing on your toes becomes table stakes.' That's what happened with spending 40% of dollars on Meta and Google.
Use DTC as a feedback loop: run small to medium-sized ad tests to key markets and test different product-level messaging, brand-level messaging, flavors per geo, and recipes to showcase consumption. By the end, you'll know which messaging works best for field marketing, retail shelf strips, social media, email blasts, and salespeople.
Scroll publisher Facebook page posts to find articles with high likes AND almost a 1:1 ratio of likes to shares. Run a post-engager ad to a prospecting audience. By the time the posts go live, they'll have a ton of engagement AND you're building a great page-engagers audience to retarget.
The beauty of running ads to a pre-launch landing page: you can set the conversion objective as email/number submission and understand what types of headlines, imagery, video, ad copy, and targeting have the highest probability of converting when you launch. These insights can inform your homepage and product page copy.
Out-of-Home can be great in condensed markets (LA, NY, Austin, SF, Miami, Boston) when you track measurement through mobile device IDs to see who made a purchase. You can also collect those device IDs and retarget on paid social.
Direct mail has increasingly become more valuable and ROI-positive for a lot of my portfolio companies. Even if someone wants to throw out their mail, they still have to skim through it.
Always question the strategy with reporting. Do you really need 10M impressions if the CTR is 0.43%? Even if the CPM is cheaper? How would it look if you paid higher CPMs instead?
You can't just keep hammering on the same audience, value props, and pitches since you started. Learn how your product is valuable to different groups and test a variety of audiences.
Use vanity URLs like nik.co/LP — the URL looks clean and memorable, but when you visit it, the UTM parameters expand at the top so you can identify the traffic in analytics.
You can buy direct mail segments like 'Whole Foods shoppers who pay with AMEX' or 'Amazon shoppers who spend more than $500 per month' from credit card data providers and send them specific offers.
Meta and Google are the best channels to start with because they have the fastest feedback loops, they're the fastest to learn and optimize, and they're the easiest to set up. You can start with ~$2,500/day and get insights quickly across 2-3 audiences.
There's no such thing as 'ads not working.' You're either running ads with bad creative, messaging, targeting, or landing pages — or your product isn't valuable enough to demand the price you're trying to charge.
I love ads because they guarantee reach. Even spending $50-100 per day on Meta or TikTok can give you really meaningful insights for your brand.
OOH is a relevant channel for brands who have found success on traditional digital channels and want to find other smart places to invest in brand awareness. Focus on memorable moments, produce eye-catching creative, get smart about repurposing the campaigns, and spend with ROI in mind.
One key trick with publisher whitelisting: make sure you reply to every comment as the publisher account. Just like you'd never let comments go unanswered on your brand ads, you must do the same with publisher accounts.
To test TV ads, you need at least $7–10M/year in revenue, battle-tested acquisition funnels, and a $150K test budget you're OK losing. Split it: $80–100K for media buying, $10–30K for creative, and $20–50K for an external channel expert.
Start with linear remnant TV ads — the less popular, less-in-demand ad slots. A $25K budget with $10–25 CPMs doesn't give you enough reach to understand if TV is viable. Once you find angles that work, layer in streaming (OTT) at ~25% of your total TV budget.
For months and quarters after you run TV, it's very common for brands to see their CPA decrease on existing digital ad channels. This is the 'halo effect' — second-order efficiencies to your Meta/Google/TikTok spend.
Customer post-purchase surveys are the absolute best way to measure lower-funnel performance. Your GA data might say someone came from a branded search ad, but the survey will tell you they were prompted by the TV ad. With KnoCommerce, we get about 67% completion rate with 7 questions.
I call this concept 'Secondary Presence.' It's using first-party data, insights, targeting, learnings, ad spend, and creative with third-party social proof, trust, validation, and storytelling. Instead of videos from influencers, it's articles from internet publishers run as ads.
In the feed, when you have an advertorial as an ad, it feels natural and contextually native to the feed being scrolled. It doesn't come off as an ad. You need to disclose the brand in the Paid Partnership tag, but as long as it looks natural everywhere else, you're good.
A publisher doesn't need to be Vanity Fair or PopSugar — it can be any editorial website. I've used 'Tier 1' publishers as well as sites you've never heard of. They work equally the same. It all comes back to the UX and cleanliness of the publisher itself.
With Pinterest traffic, we saw about a 5-minute time on site and quite cheap CPCs compared to Meta. The scale was less, but you can absolutely get efficient traffic. Pinterest ads are underrated.
If your click-through rate is below 2%, you have a creative problem.
If your CPM is above $20, you may have a creative problem. High AOV products can get away with higher CPMs, but good creative will generally still lead to lower CPMs.
You should spend up to $1,000/day per million people in your audience. So if you have an audience size of 20M people, you can safely spend up to $20K/day before you begin to saturate.
If you are spending less than $100k/month, only focus on Meta. Win there first, then expand.
If you're spending $500k/month, you should run a test on TV to allow your lower-funnel performance marketing to be more effective.
Add branded search as your second paid channel. When you're getting tons of reach on Meta, consumers will search for your product on Google. Protect your turf. Spend 2-10% of your Meta budget on Google branded search.
When spending $1k-20k/day on ads, the number one thing you should be doing outside of making the best creative possible is comment moderation. Someone hating for no reason? Hide, block, delete — in that order.
Respond to every positive comment or question on your ads. This is by far the easiest way to lower your CPMs and keep your ads juiced for a long time. Customers and prospective buyers LOVE to see responses from the brand.
Think of scaling paid media like stretching an image in Google Docs. If you rush to paid media too early before building a foundation, it gets blurry fast. But a brand with its foundation set can keep stretching.
The angle is probably the main thing people forget in advertising. Brands just starting out have zero chance at getting market share unless they have a great angle to go out with.
The audience directly correlates to the problem you're addressing in your angle. When we scaled Cadence's ads, the travel audience had a TSA-approved angle, versus a mid-20's audience which had an angle about making sleepovers at your partner's place easier.
Make ad variants for custom audiences. If targeting both 24-30 year old men AND women in NYC, test two variants — one targeted at men, another tailored for women. The more specific the better.
For cold-traffic prospecting ads at large scale, you need to over-explain the benefits and value props. This audience has never heard of you — simplifying the product and reiterating all key value props is a must.
If your brand gets written about in Forbes, Vogue, or Town and Country, run that article as an ad. The easiest way to make the most out of earned media is amplifying it through a paid campaign. You can also screenshot the headline and run it to a landing page.
You don't need a $100k piece of creative to start running TV. You just need to make sure you have a good story, a good hook, and a good offer.
Look at your top 5 ad creatives in Meta over the last 6-12 months. What can you learn about the assets, the angle, the copy, the colors, or the call to action that might work for TV? Use these ads as your starting point for TV concepts.
Fabletics tested 15 creative concepts when they first started with TV — high-production, low-production, direct response, brand awareness. After 2-4 weeks of data, they doubled down on what worked. They also repurposed Meta ads for TV and only changed the editing.
As brands exhaust their paid social audiences, TV is going to be the best place to diversify. In 2023, 88% of U.S. households owned at least one internet-connected TV device — over 110 million CTV users.
National prime time TV spots in the U.S. have CPMs ranging from $25 to $50. Non-prime time national spots range $10 to $20. Local TV spots can be as low as $5 depending on market and demo.
The best way to understand Facebook ads is thinking of it as a person, not a computer. Facebook knows its most valuable users to advertisers. Those likely to purchase see purchase-intent ads. Those who just scroll see awareness ads from advertisers wasting money on awareness campaigns.
When you run awareness or website clicks campaigns on Facebook, you're not only wasting money — you're targeting unqualified prospects. 'The CPMs are so much lower when we run awareness' is a trap. If you're not doing $1B in revenue, you don't need awareness campaigns.
Experian — the company you use to check your credit — makes most of their money selling you in a group of anonymized users to advertisers like me. I'd imagine about 30% of their revenue comes from this.
Use a redirected vanity URL behind your ads, like longwknd.com/ad1, so when BFCM rolls around and the ad has amazing social proof, you can just change the redirect destination without touching the ad itself.
CPM tells you if Meta likes your content — low CPM means Meta likes it. Click-through rate tells you if users like your content — low CTR means your creative isn't getting people interested.
Ask yourself: if you absolutely had to spend another $500K this month, which product, offer, channel, and creative would add the most net-profit — not just revenue — with every new dollar spent? How do you know this and what math do you have to back it up?
Ben Yahalom encouraged his team to 'Name the number' — make a firm prediction like, 'I believe that if we spend another $100,000 on Meta on this specific offer before 2pm, it will yield another $150,000 in profit.' It trains you to get better at forecasting, which is the hardest and most valuable skill in business.
Instead of targeting everyone, try 40-60-year-old men in a specific geo with a specific offer and matching landing page. The more specific across the TRACE funnel, the higher your CVR.
If you're spending $50-100K on Meta efficiently, test Taboola with $10K. You can typically find 4-7 publishers that respond well and keep those as always-on campaigns with steady results.
If X% of customers who buy lower-cost Product A also buy higher-cost Product B within 90 days, Product A becomes a loss leader. An inexperienced marketer might turn off those ads before the blended contribution margin proves profitable.
When I joined Hint, they were paying $4–5 just for a click to the website. Out of those visitors, maybe 2–3% converted, netting an $80–100 cost of customer acquisition. The website had a horrible UX. That's what happens when your funnel has holes.
The Hint article on The Hustle demonstrated an incredibly easy way to build a full-funnel campaign. People who had no idea about the brand could hear the founder's story, understand why she created the product, and try it with an offer — all in one piece of content.
After the Hint article broke The Hustle's website multiple times, I tried the same strategy on BuzzFeed, Refinery29, etc. Not even close. Then Digg.com took off. The secret sauce wasn't the publisher's logo — it was the story, the formatting, the length, the font, and the ad units driving traffic.
When we found ad copy that worked — customer reviews, earned media quotes, that first 'aha' moment — we pushed it to all new video tests. That way we isolated creative performance by pairing proven copy with new visuals.
When the Hint advertorial was ready, I created 96 different ad combinations with AdEspresso. Overnight, acquisition cost went from $90 to $25, and we went from double-digit new customers per day to thousands. The content was the unlock.
In an environment where CPMs are only going up, sharing engaging stories keeps your CPMs low. Native ads can be a massive unlock and more brands should try it.
If someone clicks an ad for a blue blanket, they should see the blue blanket — not a pink pillow. Ad-to-landing-page mismatch is a HUGE reason people bounce.
Every comment on your ads should get a reply. For negative comments: HBD — hide, block, delete. In that order, or they come back.
Catalog ads don't make sense with 1–2 SKUs. But with 20 to 100+ SKUs, they're incredible. You can even make a smaller catalog of ~10 products work with enriched creative overlays.
The future of advertising keeps getting more automated. Meta is pushing Advantage+ Shopping Campaigns hard. It's possible they're quietly juicing catalog ad performance to get more brands using them.
According to the Marpipe x Varos report on 3,500 companies spending $4B on Meta: fashion brands saw 50% higher ROAS and 40% higher CTRs with catalog ads versus traditional static or video ads.
The point of diminishing returns is when you're spending more than a channel can support, causing CPA to rise for no reason. At a previous brand, we lost a few hundred thousand dollars learning our ceiling was $1.5M/month on Meta. Media mix modeling would have caught it.
With tracking losses, all content should feel full-funnel. Whether someone is top or bottom of funnel, every ad and landing page should educate and sell. You never know which user wasn't properly tracked and got served the wrong thing.
Most top-of-funnel ad dollars should still go to Meta and Google, with maybe 20–25% toward YouTube. The key to standing out isn't campaign setup — it's all in the creative.
CPMs on remnant TV can be as low as $3. Convergent TV inventory still massively amplifies your performance paid social. Don't sleep on TV.
Start TV simple — one engaging 60-second founder story. Think of it like paving a road with reach. The Ferrari driving on that road is your performance ads. The better the road, the smoother the ride.
A reliable native ads funnel: drive traffic to an advertorial you know converts, test across publishers. You'll find 3–10 that work well. Double down on the winners.
A common mistake: wanting to be everywhere at once — Meta, TikTok, YouTube, Google, Pinterest, Snap, podcasts. Moiz reminded me: Meta alone can scale a company past $10M/year. Master one channel first.
Moiz puts it this way: When you pay Meta for ads, you're paying for Malibu real estate. Would you let someone spray-paint your garage saying your brand sucks? Protect your ad comments. Hide, block, delete — in that order — and respond to every positive one.
Nood spent $500K last month on TikTok with a platform-reported ROAS of 0.5x. According to TikTok, they lost $250K. But using a multi-touch attribution model, the channel was actually profitable. You can't always trust what the platform tells you.
You should use different attribution models for different channels. Each platform has slightly different user behaviors and consideration periods.
Bryan at Nood is seeing $2-3 CPMs for linear TV on various networks versus Meta, which could be a $20-30 CPM. This is the perfect example of arbitrage and underpriced attention.
When you run awareness ads on Meta, my hypothesis is they serve this traffic to the lowest-quality users — the folks who never click or engage. You can get infinite reach and incredible CPMs, but the site traffic and purchase behavior is going to suck.
Even before Meta was having issues, brands were getting 10-200% higher ROAS after enabling catalog ads. Meta is clearly pushing this ad format and the performance has been undeniably strong.
Have two or three ad accounts for your brand just in case one has issues and you need to migrate spend within a day.
Snap media is still underpriced. CPMs are competitive and people aren't paying attention because TikTok is the shiny new platform.
The beauty of TV creative: once something starts working, you can run it efficiently for 3 years in a row with little to no change in performance.
Programmatic ads only make sense coupled with another channel, usually Meta — as a reminder/retargeting effect. Watch the attribution window. Programmatic over-attributes in my experience.
Host-read podcast ads drive higher ROAS. You're not just buying time — you're getting the co-sign and influence. When you find the right podcast x brand match, podcast ads are still underpriced.
Think of Meta's catalog ads auction like an Express Lane — less competition and everyone's creative is boring. Make your catalog ads stand out and you're in the carpool lane of Meta.
Scope out competitor ad libraries to ensure you're not missing anything — and to make sure they're not making stupid claims against your brand. So many 'loser brands' run ads taking shots at the category leader. Keep an eye on them and send a C&D while they continue to lie.
Beware of sneaky ASC+ creative enhancements on Meta. There's nothing worse than cut-off ads for a high-quality brand during peak shopping time. Meta is trying to create as much weird ad inventory to keep us spending — don't get suckered into a bad spot when you can avoid it.
If it's a digital channel — YouTube, programmatic ads, Taboola — I aim for a test budget of $10-15k. With digital and performance-based channels, you can always cut the test if it's not looking good after $15k.
For publishers, minimums usually start around $25-50k. For podcasts, a good test budget is probably around $50k — you want to test multiple shows with different messaging.
I am a fan of using agencies to test a new channel. Agencies can be faster with the 0-1 of a new channel, have more creative capabilities, and can crack the code before you scale it. Then bring it in-house once they find scale.
The beauty of digital ad channels is that almost everything can be tested fairly quickly. Within 2-4 weeks, you will understand how well a new digital channel may perform. If you're running direct mail, out-of-home, or TV, you need a longer window to judge performance.
A viable YouTube or Podcast test should be around $30-50k. The beauty of YouTube organic is that the URL stays in the video forever. If the video continues to get views — which they do — you'll be generating purchases forever.
When launching new channels, the most important thing is to track the incremental lift from each channel. Every platform will over- or under-report its numbers. 65% of the time, in-platform numbers are overreported — which is why I'm a fan of understanding incremental lift.
TV is cheaper than you think to test. Remnant ads can be $1-3 CPM to get on a prime network like ESPN or CNBC. Meanwhile, on Connected TV, you're more likely to pay $10-30 CPM, but you're targeting the user, not the content.
Do not turn off evergreen ads that perform well during the holiday season. Instead, add holiday-specific ads on top of your existing evergreen campaigns.
In October, allocate 10% of your marketing budget to run video view campaigns on TikTok, Facebook, and YouTube focused on ownership benefits. Then in early November, run lead gen to build an early bird list. This list can drive 50% of BFCM sales.
DPAs (Dynamic Product Ads), especially enriched catalog ads, often deliver up to 2x better ROAS vs standard statics or videos. Catalog ads are seeing increased ad spend while video ads are declining. The time to lean in is now.
During an election year, customer acquisition costs rise dramatically on Meta and TV in swing states due to political ad spend. Target states with less political ad spend to find pockets of more affordable advertising.
Snapchat's fastest-growing demographic is 25+, with nearly half of all U.S. Snapchatters in this age group. 43% of U.S. Snapchatters are not on TikTok daily — that's real incremental reach.
Meta DPAs are performing better than almost any other ad format right now. Many brands are seeing up to 2X the ROAS compared to traditional static and video ads. My guess is DPAs will be the top performing channel by total revenue for most brands.
The creative IS the targeting now. If you show an elderly man in your ad creative, Meta will find users who look like that elderly man. AI-enabled ad algorithms like ASC, Pmax, and Smart campaigns mean most targeting is done by AI — your creative does the work your media buyer used to do.
CPMs on TV are low, they bring validity to your brand — not just anyone is on TV — and once you find a winning creative, you can run it for months. A few years ago, running CTV ads required a specialized agency and tens of thousands just to test. Now you can start with a $500 budget.
Advertorials now account for roughly 22% of ad spend on Meta and help balance out the new customer CPA. Your favorite performance marketing brands are all leveraging sponsored editorial content, leading with entertainment and then educating customers on their product.
An advertorial is just a combination of an ad and editorial piece. Story first, sales after. In 2017, I started running advertorials and whitelisting publishers and creators, and it blew the brand up more than anything else. Eight years later, this strategy still works.
When we found an ad angle for diet soda drinkers at Hint, it generated 8 figures in revenue. Testing new messaging angles with large TAMs can be the biggest lever for scaling.
There's a direct correlation between CTR being up, CPM being down, and CPA being down when you're constantly replying to every single ad comment on Facebook and Instagram.
A gym bro in Wyoming wants colostrum for different reasons than a 35-year-old woman in Manhattan. You can't find these individuals without testing new Audiences, Messaging, and Offers — AMO testing.
In 2017 when I started using incrementality for measurement, it illuminated that Snap drives real retail sales, programmatic ads do absolutely nothing for DTC sales, and even what type of creative would drive sell-through in Target.
On Meta it's ASC. On Google, it's Performance Max. On TikTok, it's Smart+ campaigns. In all of these, your ad creative and the contextual clues within it become the targeting. There are rumors Meta is going to take away targeting altogether.
On TikTok, a verified creator or influencer handle has a CPC usually 50% lower. The CPM of whitelisted ads is generally lower than ads coming directly from the brand's own pages.
For DTC to work well, all the marketing mechanics need to work together under one roof: acquisition, retention, and commerce experience. If they're in unison and can move quickly, then Facebook ads just 'work.'
Fixing Facebook ads has everything to do with everything outside of Facebook Ads Manager itself — measurement, ad creative, hooks, the problem you're solving, landing pages, PR, and testing new organic channels.
The cost of advertising is going up. Meta and Google CPMs are steadily rising. You can't just buy customers anymore; you have to earn attention across the funnel. This starts with brand perception and ends with checkout UX.
When people sit down on the toilet and open Instagram, no one thinks, 'I can't wait to scroll down a brand's PDP.' People want to be entertained. That's where advertorials shine. Advertorial = ad + editorial.
Your conversion path should be: Meta Ad → Advertorial → Optimized PDP. It warms up the click and filters intent before you pay for a bounce. The CPC to an advertorial is 90% lower than going to a PDP.
We've seen 30-40% reductions in CPA from advertorials, along with strong halo effects in retail and Amazon from incrementality studies.
On average, CPCs with whitelisted handles are 25-35% lower on Meta and 30-55% lower on TikTok. At the very least, run ads using the founder's Instagram.
Video Catalog Ads (PLV) are now outperforming static by 30-40%.
A channel like YouTube may not report significant numbers in-platform, but an incrementality study will show it drove many new customers over 8 weeks. Run hold-out tests before killing channels.
Front-load information — begin ads with the strongest, most compelling benefit. Someone won't spend more than a second or two, so keep the important stuff up front. I do this with subject lines, too.
When running any view-through channel — influencer, TV, podcast, radio, billboards, sampling — always use a vanity URL, coupon code, and post-purchase survey. This gives you the best chance of triangulating attribution.
If you're not running spark ads on TikTok, you should be. CPCs with verified creator accounts are half. You're throwing money in the garbage if you're not running spark ads on TikTok.
The goal of your ad isn't to convert someone — you just need someone to click it and get to the next page. It's on your homepage, PDP, or landing page to convert the user.
When running top-of-funnel with small budgets, focus on TOF channels that can be run like performance channels (TV, advertorials) before expanding to those that can't be measured as easily (OOH, taxi-tops, sponsoring events).
Static ad creative CPMs on Meta are about one-third the CPM of video ads. Turn your angles into statics in Figma and test them alongside video.
Don't go broad on Meta with $0 in historical spend — despite what you see on Twitter. Going broad works when you have enough history for Meta to know who to target. Think of the Meta ads engine as a human: if it has no learnings, how will it know what to do?
High CTR but low conversion rate? Your site needs work. Low CTR? Your ads need work. CTR is the diagnostic that tells you where your funnel is broken.
If CPM is high on Meta, the ads you're uploading don't match the audience or aren't friendly to what Meta wants in their feeds. You're being overcharged for putting something the platform doesn't want.
If Add to Carts are low but CPM and CTR are strong, the problem you're solving is real in the ad but isn't consistent with the website. You need to test landing pages or better site pages.
Partnership ads on Meta can acquire customers for 20-30% lower CPA. For scaling brands, they only represent 20-30% of media spend — so much room to scale without diminishing returns.
If you see Gruns running a new landing page test with a funnel for moms sending their kids back to school, and want to run one, but it requires a budget approval to hire an agency to build a landing page and another approval to get $5k in ad budget to test the audience, you'll be 4 months late. By that time, that angle is DONE.
The more data you feed IN to the ad platform, the more you can get OUT of them. If you launch without having the proper pixel setup across your website, you're missing out on training data for your ad account.
If you're acquiring cheap customers who aren't good customers — they buy once, never come back, low subscription rate, cause chargebacks — it's not worth it. You have to understand if you're driving quality customers, especially if you're building a brand for the long term.
If you don't build more top-of-funnel awareness, you can't continue to scale your performance ads efficiently. That's why the brand CPA keeps going up. You're reaching fewer people who know your brand, conversion rate tanks, and acquisition costs go up.
Channels like TikTok, Snap, AppLovin, TV, organic YouTube integrations, podcasts don't just 'not work' — they don't work because you likely didn't test them properly. If your product delivers on its promise, any channel that reaches your audience can work.
I've been screaming about whitelisting for almost 9 years now and brands are still hesitant to try it. Whether you do it with advertorials, creators, meme pages, or just the founder's own Instagram — try it. If any of these pages are verified, you'll have EVEN better results.
The best brands I know don't wake up and get mad when their CPMs go up. Instead, they take their budget from one channel and move it to another because they have better CPMs that day. They can only do that if they have funnels, offers, creatives, and messaging that work for each channel.
Start new channel tests with retargeting to leverage a shorter feedback loop. Like a hawk, look for where the traffic leak is and plug it using heat maps and session recordings. As you find your funnel clicking, move the campaigns into colder audiences.
Partnership/whitelisting ads ALWAYS outperform your evergreen, business-as-usual ads. I've seen a 2x better CTR when brands leverage a verified creator to drive ad traffic. No one wants to hear from companies — they want to hear from people.
I got a creator's Instagram and Facebook pages in our Business Manager and lit up the ads. The CMO said, 'TAKE THIS DOWN, THIS IS A TERRIBLE LOOK FOR THE BRAND.' I said, 'It has the lowest CPA and broke us into a new audience that's scaling fast. We already spent $25k in 24 hours at the lowest CPA.' That CMO didn't last.
Streaming now accounts for 45% of all USA TV viewership — more than cable and broadcast combined. The halo effect from TV is real: it lifts Meta, Google, and branded search simultaneously. You don't need a $500K budget to test it.
As people engage with you in Facebook Messenger, those users become addressable for one year. If you don't have the budget for SMS, start with Messenger — it's a free addressable audience.
Over $10M in YouTube spend in 2020 averaged $0.18 CPV across home goods, F&B, and beauty, with an average ROAS of $8.65 — almost $90M in revenue. The split was 70% prospecting, 30% retargeting.
YouTube's Video Action Campaign (VAC) format was over-crediting itself. But after $1M in testing, CPMs came in ~85% lower than TrueView for Action — if set up correctly against the right audience, you can achieve 30-45% CPA reductions.
When COVID hit in March 2020, YouTube CPMs decreased 40-45% on desktop and 30-35% on mobile. Brands that scaled up 65% during the panic saw massive returns by capitalizing on cheaper CPMs and the surge in online shopping.
When responding to ad comments on 7-figure monthly paid social budgets, we saw acquisition costs drop by about 20% on average, sometimes more. The insights from unanswered questions also tell you exactly what to fix on your homepage, product page, or landing page.
When testing sponsored editorial content across BuzzFeed, Refinery29, PureWow, Babe, Digg, Forbes, and The Hustle — my guess was Refinery29 and BuzzFeed would win. The winners were actually Digg and The Hustle. Why: little or no programmatic ads to distract readers, easy-to-read fonts, non-distracting pop-ups, and clean mobile experience.
If you've built your brand completely organically but don't know what audiences work best on Facebook, just go broad with your targeting and let Facebook's machine learning figure it out. As long as you've had a Facebook pixel on your website, Facebook has secretly been learning who your customer might be.
The 3 traits of a great performance PR article: 1) It tells a story — the editorial 'I tried this and...' angle always works for introducing new products. 2) It isn't cluttered with pop-ups and programmatic ads. 3) It really describes the product so you almost feel what it's like.
If you see a PR article generating sales right away — through Google Analytics or affiliate links — and a higher conversion rate from readers, that's your signal to amplify it with paid. That article just became your best ad creative.
People dismiss TikTok ads with objections like 'the pixel is only last-click,' 'traffic only stays 10 seconds,' or 'the audience is too young.' The real problem isn't the platform — it's that people run ads on new platforms using the same playbook and measurement framework as Facebook, then blame the channel when it doesn't match.
The unlock with Performance Branding is your ads become more 'invisible' as ads because the content is so relevant. At Hint, some content had a 13% click-through rate and got the account shut down for falsely tripping the system from being too high.
If you add a swipe up to Snapchat, make it easy. Don't make it your homepage, and don't make it the same landing page you would send Facebook/Google traffic to. It should have 1-3 options of where to click, or it should be email/phone collection to make it a quick experience.
A positive sign that Snapchat retargeting is working: your time to purchase gets lowered when adding in the channel. An LVMH brand with a $2,000 AOV said Snapchat retargeting was their most efficient channel — look for view-through conversions, not last-click.
Are you segmenting retargeting campaigns by funnel step — site visit, product view, add to cart, begin checkout — and pushing different messaging at each stage? This should also match your site-abandon email flows.
Are you cutting up your customer list properly and creating targeted lookalikes? High LTV customers, high NPS customers, fastest subscribers, longest video-viewers, high AOV — each should be its own lookalike seed.
Does the AOV you're pushing match the paid channel a user comes from? Each channel has different user intent and spending behavior — align your offer's price point to the channel.
Each ad channel reacts differently to price points. Meta and Google are the most unbothered by AOV. TikTok and Snap do better with lower AOVs. Pinterest is in the middle. Match your offer to the channel's price tolerance.
You could acquire customers from TapJoy at a fixed CPA all day, but then realize none of them come back, they initiate chargebacks, and they clog your customer service inbox. The source of your customers matters as much as the cost of acquiring them.
A form of paid media that's rarely done but can be effective with large budgets is buying display inventory on premium publishers directly. If you can get an article coupled with that, even better.
Pre-iOS 14.5, Facebook knew everything — how long you spent on a site, what products you looked at, which reviews you read. Driving traffic to a mediocre homepage could still work. Today, ad platforms don't get any of that data, so you need to create the reason someone would buy from ad creative through the purchase experience.
The 2017 playbook of uploading customer lists to Facebook, loading remarketing campaigns, and hitting publish could guarantee purchases at a few dollars each. Running that same playbook today, the CPO comes to roughly the same as acquiring a new customer. Single-channel marketing is dead.
Buying the Dip: August CPMs drop on ad platforms — spend money then to build higher-intent audiences you can retarget for Black Friday, Christmas, and the holidays. You're essentially buying cheap awareness now and cashing it in during Q4.
Channels where CPMs DON'T inflate during Q4: trackside ads (truck wraps via Adgile), direct mail (paper, postage, and ink costs don't rise just because P&G deploys money), and affiliates (they don't charge higher percentages during peak season). Use these for short-term arbitrage when paid social CPMs spike.
Set up conversion campaigns in Meta pointed at a microsite to test how different creative and messaging produce different results for post engagement, CTR, CVR, and ROAS across local markets. Use those insights for field marketing, retail shelf strips, social media, email blasts to local regions, and sales conversations.
To build a highly qualified retargeting audience for $50-75/day: find publisher Facebook pages your customers follow, scroll for posts with a near 1:1 likes-to-shares ratio, schedule those articles as posts on YOUR page, then run a post-engager ad against the scheduled post ID. By the time it goes live, it has engagement AND you've built a page-engagers audience to retarget with your BFCM offer.
If you're running paid ads through a URL redirector to avoid re-entering the learning phase, triple-check your redirects on multiple devices and incognito windows. A broken redirect during BFCM is money on fire.
Pre-iOS 14.5, I could use AdEspresso to launch 48 versions of creative at $5k/day and within 72 hours know which 5 would scale to $30k/day. That playbook is dead. Now every ad needs to be full-funnel capable — it has to work whether someone knows everything about your brand or has never heard of you.
On Meta ads, always attach a conversion event to your video ad — email signup, store locator click, or a lead gen form. When you set a proper conversion goal versus running for reach or engagement, your ads get served to a much higher-quality audience. The goal of the creative should always be a click.
Using ad platforms the same way large companies do doesn't work for brands under $100M in revenue. Uber can run a World Cup TV spot about delivery times without explaining what Uber Eats does. For DTC brands, every ad needs to do the heavy lifting of educating AND converting.
As of early 2023, YoY channel budget data from Northbeam showed: Meta spend down 10% with CPM down 18.7%, Google spend up 17% with CPM up 15.9%, TikTok spend up 32% with CPM up 12.8%, YouTube spend up 2% with CPM down 40.5%, Microsoft (Bing) spend up 113% with CPM down 35.9%, Pinterest spend up 501% with CPM down 28.3%, and Snap spend down 56% with CPM up 33.5%.
Over 150 million residences across the country receive mail — that's homes, not individuals. With 2.6 people per household, your potential direct mail audience is roughly 390 million. Facebook hopes to have 243 million people in the USA by end of 2023.
When you start out using direct mail, start with an audience already familiar with you. Retargeting has a higher engagement benchmark, so you can validate messaging, creative, and site experience before opening up to prospecting.
When you're just getting started with Meta ads, use a manual ads manager setup. Your pixel isn't well-trained yet, so you can't expect it to optimize perfectly with Meta's more automatic campaign setups.
Don't even think about turning another ad channel on until you can profitably spend $5K/day on Meta. They have the fastest feedback loops, are the easiest to set up, and you can start with ~$2,500/day to get insights across 2-3 audiences.
Pixels lie, attribution software can lie. You know what doesn't? Good old pivot tables, LTV cohort analysis, and MER. Ben Yahalom at True Classic let data dictate every decision — contribution margin by channel, payback period for net new customers, Meta allocation on good and bad months.
About 15% of consumers are opted-in to sharing mobile device location data through free apps. AdQuick uses that to build attribution models around OOH campaigns — you can see if someone who walked by your billboard later purchased on your website.
Whitelisting works 10x better in B2B marketing, because B2B is like 10 years behind DTC. The same secondary presence tactics that crushed it for consumer brands are barely being used in B2B.
The $150k TV test budget breakdown: $80-100k for media buying (enough to test placements, networks, and TV media types), $10-30k for creative production (up to $100k for high-quality), and $20-50k for an external channel sniper who knows TV inside and out.
With advertorials, you must disclose the brand in the Paid Partnership tag or you'll get your account banned permanently. But as long as the article title, image, and caption feel natural like editorial content, you're good.
When you run an Awareness campaign on Facebook, you're telling the algorithm to find people who will see your ad — not people who will buy. Facebook knows who buys and who just scrolls. An Awareness campaign literally optimizes for people who won't convert. You're paying to show ads to the cheapest eyeballs, not the most valuable ones.
Think of the Facebook ads engine as a person, not software. I call him 'Ethan.' When you set a campaign to Awareness, you're telling Ethan to find people who are good at seeing things. When you set it to Purchase, you're telling Ethan to find people who are good at buying things. The objective you choose is literally the instruction you give Ethan.
The 5 things that actually matter in paid media: the angle, the audience/targeting, the ad creative, the web experience, and the offer/merchandising. That's it. You don't need 46 SaaS apps and 4 ad account strategies to test against each other.
Fabletics has repurposed Meta ads for TV by only changing the editing — and it performed well. You can start TV with $10K in production and $5-10K/week in placements. You don't need a $200K creative budget.
Once you've profitably spent $250K/month on digital ads for 1-2 quarters, you're ready for TV. At Hint, we didn't test TV until spending $1M/month on digital. At Native, Moiz was at $1-2M/month. But $250K/month with proven demographics and creative angles is actually enough.
Your ads are published to something like a For You page. When someone engages with a fitness video, they see more fitness. Go deeper into cookware, and you shift from cooking to cookware. Your creative steers who sees your ad — the angle and how you speak about the product determines your targeting more than any audience setting.
Facebook's Social Connections data means your friends' likes and interests influence what ads you see. This is why you and your friends see the same TikTok videos — and why word-of-mouth creates a compounding effect in paid media too.
The Facebook Business Manager view I set up at every brand, in order: Delivery, Ad Set Name, Bid Strategy, Budget, Results, Reach, Impressions, Cost Per Result, Quality/Engagement/Conversion Rankings, Amount Spent, CPM, Link Clicks, CPC, Content Views, Cost Per View, ATCs, Cost Per ATC, Checkouts, Cost Per Checkout, Purchases, Cost Per Purchase. If any right-side columns are blank, you're running the wrong campaign objective.
True Classic's Ben Yahalom had his team meet every two hours on Black Friday starting at 6 AM. Each person had to 'name the number' — make a firm prediction like 'If we spend another $100K on Meta before 2pm, it will yield $150K in profit.' Then they measured results and held themselves accountable.
Ben Yahalom pulled back on BFCM spend after naming a number and seeing results didn't match his prediction. He didn't chase the 'BFCM opportunity' — he stood by efficiency and profitable spend over top-line growth.
During election years, YouTube CPVs can spike from $0.06 to $1.50 per view in key states around major political events. Facebook CPMs rise 15-30%, and TV ad rates increase 10-20% nationally and up to 50% in battleground states.
Use Google's Transparency Reporting Tool during election years to see who's spending on political ads, how much, and in which zip codes. It's a masterclass in audience segmentation — you can see how candidates cater creative to different demographics.
At Hint, I was paying $4-5 per click to a site converting at 2-3%, netting $80-100 CAC. I paid a few thousand dollars for a founder story article on The Hustle, ran 96 ad combinations, and overnight our CAC dropped from $90 to $25. We went from double-digit new customers per day to thousands.
The secret sauce of the Hint advertorial wasn't the publisher's logo — it was the story, the formatting, reading length (2-3 minutes), what else was on the page, the font size, and the ad units driving traffic. BuzzFeed and Refinery29 flopped. Digg.com worked. In 2018, I drove 1/3 of all Digg's traffic with the same playbook.
With the Hint founder story, the content was so good it got hundreds of thousands of organic shares on Facebook. The wholesale business took off as a byproduct — people walked into retail stores saying they read the story on Facebook. Great content-driven acquisition creates a halo effect across all channels.
If ad platforms offer partner or accelerator programs like Facebook's old Disruptor program, do whatever it takes to get invited. When Hint joined, it was the cheat code — the Disruptor team could help any brand scale. These programs give you insider access to how the algorithm actually works.
When Hint was paying $4-5 CPCs on Facebook with a 0.5-1% conversion rate, that equated to an $80-100 CAC. I applied clickbait-style native advertising tactics to get CPCs down to 5-10 cents — that's what unlocked the entire business.
Native advertising spend was projected at $97BN in 2023 and expected to rise to over $300BN per year by 2031. It's a $100BN/year ad category that most DTC brands aren't leveraging at all.
The clickbait native ad model: get a 1-2 cent click with Taboola or Outbrain, send them through a 10-30 slide article, and if they get 4-6 slides in you make 5 cents and profit 3 cents per user. At scale, hundreds of thousands per day, those pennies add up fast.
For catalog ads on Meta, select 'Catalog Sales' as your campaign objective and run it as an Advantage+ campaign. I call it Chad Targeting — surrender to the algorithmic Gods. Meta is pretty good at optimizing on its own.
Nood spent $500K/month on TikTok with a platform-reported ROAS of 0.5x — but after building a custom MTA model in Triple Whale, they found the channel was actually profitable. TikTok's native metrics don't tell the full story.
Bryan at Nood uses 1-day click conversions as a proxy for how today is performing and 7-day click conversions as a lagging indicator. There have been multiple occasions where 1-day click metrics looked bad, but 7-day click looked awesome — so he kept spending.
Nood's consideration-to-purchase period is only 7 days. Every brand should measure their own consideration window per channel — TikTok, Meta, and Snap each have different user behaviors and lag times between first touch and first purchase.
When Meta is broken, catalog ads are your carpool lane. Even before Meta's issues, brands were seeing 10-200% higher ROAS after enabling catalog ads. Meta is clearly pushing this format — master it.
If something is really wrong with your Meta ads, try creating a new ad account, a new pixel, and a completely new payment method. But be careful — immediately creating multiple ad accounts and running different credit cards can flag your account for disabling.
CPMs on Meta were up 5-50% for many brands since March 2024. The weird thing: not all brands were affected. Some saw business as usual while others were fighting to stay profitable. Sometimes the more you do during platform issues, the worse it gets.
Running ads is a drug. Before you spend a dollar, know your COGS, shipping costs, pick/pack/storage fees, and leftover margin. From those numbers, derive your allowable CPA. The OGs say CPA, the noobs say CAC — I don't make the rules.
Don't test into a new channel just because your founder came back from a conference and heard another brand is doing it. That's always a terrible reason. Confirm the channel has your audience, the scale you need, and that YOU have the resources to properly test — both media costs and creative costs.
When launching a new channel, take funnels, offers, and messaging that already work on existing channels and amend them for the new one. They act as a control — you already know they convert, so you're isolating the channel variable.
Use iSpot TV to look up TV commercials and judge their performance based on how long they've been airing. Use AdBeat to spy on native and programmatic ads. For digital channels, check their Ads Libraries to see competitor creative and funnels.
Adding new product creative into an already-learned ad campaign will reinvigorate its performance. Don't launch a separate campaign for every new SKU — inject it into what's already working.
Cody Plofker ran a YouTube incrementality test and found that click-based reporting was underrepresenting the channel's true impact. YouTube ads at normal spend drove 1.82x more orders than click attribution showed. Doubling YouTube spend resulted in 2.26x more new customer orders — proof you could confidently scale the channel.
47% of all U.S. TV streaming time happens on Roku. They now let brands run self-service CTV ads for as little as $500, and Shopify merchants can launch shoppable ads where consumers check out on-screen using their Roku remote. The process that used to take weeks now takes less than one day.
CTV ad spending hit $30 billion this year, up from $26 billion last year, and analysts project it could surpass $40 billion by 2026. It's one of the fastest-growing ad channels in the world right now.
Roku users streamed 286 million hours of content on Black Friday — an 18% jump over normal viewing. The Sunday between Black Friday and Cyber Monday has been Roku's top streaming day for three years running. Roku powers 47% of TV streaming time in America.
Roku's Action Ads let viewers take action directly from their remote — and it integrates directly with Shopify Checkout, so consumers can check out on-screen using their Roku remote. Self-service CTV campaigns that work like digital ad platforms.
Meta is restricting health and wellness advertising across three sub-categories: patient portals with sensitive data, health providers offering telehealth, and websites catering to specific medical conditions. Upper-funnel events above Add to Cart remain; mid and lower-funnel events get restricted.
Meta calls them 'Sensitive Topics' and the signal 'Sensitive Data.' About 10 total topics are being restricted — healthcare, finance, and anything signaling hardship or conditions. If you're using an out-of-the-box pixel/CAPI solution, you're likely sending back more data than you realize.
On January 1st, all the big players' ad dollars go to zero, and you have the lowest CPMs of the year. This is the best time to start testing new channels like TV, AppLovin, or advertorials. CPMs double or even triple in Q4 — January is when you experiment.
AppLovin is all the rage for brands right now. If you have significant scale on Meta, it's worth a test — but it's not as efficient for new customer orders yet. That can change quickly if they add exclusion capability.
Video DPAs are going to be huge in 2025. Meta, Snap, and TikTok are all going to leverage video DPAs much heavier — if you're not already testing video in your catalog ads, start now.
Roku partnered with Shopify to let viewers click to buy directly from their Roku remote. Once they click, Roku sends them a text with a link to checkout. You get full control over the creative, SMS messaging, and URL with UTMs — finally making CTV a direct response channel.
In 2017, I ran advertorials + whitelisted publishers and creators at a beverage company. When we were spending $1-2M/month on the articles, people would walk into our San Francisco office and tell us they just read a cool story about the founder. Advertorials aren't just ads — they become part of a consumer's primary content consumption.
99% of programmatic ad tech companies are lying to you about reporting. They use long attribution windows, view-through-based numbers, and often arbitrage inventory cost, resulting in trashy placements. If a vendor has a 90-day view-through window, they'll get credit for almost every conversion.
A 9-figure DTC brand with flat growth had an ad platform team suggest investing $300k in media behind 6 influencer videos at $10k each. Platform reps suggest strategies to brands, tell them where to place media dollars, and recommend partners — but they've usually never run a single dollar of media in their life. Don't let them drive your strategy.
Roku predicts their self-serve ads platform will bring 20,000 new-to-streaming advertisers to CTV. A few years ago, running CTV ads required a specialized agency and tens of thousands to test. Now you can start with a $500 budget on Roku and track performance like any digital channel.
Whitelisting doesn't require big influencers — it could be a founder's account, an employee's account, a meme page, or even your dog's IG. Anything except your brand name is worth trying. Whitelisted ads are typically only ~20% of spend in most Meta accounts, but they outperform because they fit the context of the feed.
Consumers now spend more time streaming TV (3.5 hours) than on social media (2.8 hours). Almost half of US TV streaming time is on Roku. If you're selling high-AOV products, TV advertising is one of the best channels to drive awareness and trust.
The lack of tracking data post-iOS 14 means what you see in ad platforms isn't real data — those are all models built by the platforms. This is why it's increasingly important to build your own bank of zero-party data and layer on MTA, MMM, and incrementality testing tools.
Secondary Presence is using first-party data, customer insights, learnings, ad targeting, and ad dollars combined with third-party audiences, trust, social proof, and content. Think of it like a math equation — your 1P assets plus their 3P credibility.
If you run dynamic product ads (DPAs) or catalog ads, all that traffic goes directly to individual product pages — yet most brands never optimize those PDPs. Your PDP IS your landing page for catalog traffic.
When running any view-through advertising channel — influencer, TV, podcast, radio, billboards, programmatic — always use a vanity URL, coupon code, AND post-purchase survey together. This gives you the best chance of triangulating attribution on channels where there's no direct click.
If you run catalog ads as part of your media buying strategy, you should be running better creative than just the product image on a white background. Enriched catalog ads with lifestyle imagery and overlays dramatically outperform default white-background product shots.
Create a custom view in Meta Ads Manager — I call it 'Nik's Performance View' — that includes: CPM, CTR, CPC, Content Views, CPCV, Add to Carts, CPATC, Initiate Checkout, and CPIC. This lets you diagnose exactly where the leaky pipes are in your funnel instead of just staring at a bad CPA number.
Fatty15 drove 97K+ page views on Roku's CTV platform, with more than 25% of cart adds turning into purchases. Energy app Arbor reached 153,000 unique households at just $0.05 per unique reach. CTV is becoming a legitimate performance channel, not just a branding play.
CubCoats scaled AppLovin from $1k/day to $80k/day in 30 days. If you're a performance-marketing-driven brand willing to test new channels, AppLovin is worth testing aggressively right now.
You can hit $50M/year in revenue with just Meta as your primary driver. Don't test a TON of channels — find a couple you can master that are also conducive to your business. Selling a mattress on TikTok Shop? Waste of time, no matter how good your TikTok Shop skills are.
Set up all ad accounts and tracking BEFORE launch. Treat your ad account like a muscle — if you don't exercise it with real signals, it won't be strong enough when you need it. Every untagged visitor is missed training data.
Use native ads for top-of-funnel awareness during BFCM. You can set up campaigns on publisher sites like USA Today and Bloomberg in a few clicks — run your BFCM sale as punchy image creative or a listicle advertorial, plus retargeting campaigns.
For partnership ads on Meta, you can run ads with no button CTA so it looks organic and entertaining. At Hint Water, we spent $25k in 24 hours at the lowest CPA running a creator's cut YouTube content as whitelisted ads. I launched 96 ads with a $5,000/day budget to test if a publisher advertorial had legs.
Roku's Action Ads are becoming a real performance channel for DTC. Customer sees your product on TV, scans a QR code or clicks with their remote, lands on your Shopify store, buys. All tracked, all measurable. With 125M+ streaming households and 76% of Roku users shopping sales online, CTV is no longer just awareness.
For catalog advertisers, segment your product sets by cultural macro-trends each year. In 2024, compact digital cameras exploded 2,345% and embossing stamps were up 1,119% — products positioned as analog alternatives to digital overload saw the strongest delivery and ROAS.
When I see brands test channels with a measly $2,500 and then determine it's unsuccessful, it makes me roll my eyes. Testing with rigor means committing real budget, real time, and real creative before deciding a channel doesn't work.
The Sharma Way to test new channels: (1) funnel research — set up a fresh account and go through competitor funnels, (2) offer creation specific to the channel, (3) creative planning native to the platform, (4) web/destination planning, (5) experiment design. No more 'we tried it but it didn't work' after a lazy 7-day test with $10k.
If 70–90% of your new customers come from two platforms, you're not diversified — you're dependent. Brands doing $5M to $25M are usually stuck on Meta and Google. The best brands shift budget between channels daily based on CPMs, but they can only do that because they've built funnels, offers, and creative specific to each channel.
Roku's Action Ads let viewers press 'OK' on their remote to send a product link to their phone or buy directly on the TV. Fatty15 used Roku's Shopify integration to add shoppable overlays and drove a 25%+ conversion rate from cart adds — setup took under 5 minutes.
96% of Roku households now see video ads, and 74% of all TV watching is ad-supported. CTV is no longer optional — it's becoming one of the most efficient ways to reach net-new audiences, especially as social and search face rising challenges like zero-click results and AI-generated content flooding feeds.
The biggest waste of time is making a performance marketing media forecast with no prior data. Don't let an agency build a forecast when you have zero historical performance to base it on.
For a Super Bowl ad, the fully loaded cost is $16-29M all-in for a 30-second spot including production, talent, and required companion spend. The downside is capped at a 1-5% hit on total marketing efficiency. But if it improves future marketing efficiency by even 2% for a company spending $500M/year, it pays for itself without a single direct conversion.
Wispr Flow ran streaming TV campaigns on Hulu, Disney+, CNN, and ESPN — QR traffic converted at 20%, with a lower blended CAC than their bottom-of-funnel Meta campaigns.
Programmatic CTV on its own is just 15-20% of total TV inventory. Judging TV performance on that alone is like running only branded search keywords and deciding paid search doesn't scale. The real risk is testing an incomplete version of TV, leaving 80% of inventory on the table.
20.8% of CTV impressions are fake content and 65% of fraud in CTV environments is bot fraud. When testing CTV, go reseller-free with direct publisher integrations and authenticated impressions to avoid the black box.