Vibe.co is now added to our dataset! We’re really excited about this emerging ad platform. (If “emerging” means “just got acquired for a billion dollars” — tells you everything you need to know.) Connected television rules, and you should be advertising there.

Obviously, because it’s connected television, you can’t track “clicks” on a TV ad. So there’s no clicks metrics in these charts.

We’ve also left out ROAS (for now) because the ROAS calculations for all other platforms in here are based on last-click. Which, again, you cannot click on a television. Not yet, anyway.

The major shifts you see in this data is because we’re comparing to July 4th. Most advertisers were blasting extremely wasteful ad spend budgets during that time to capitalize on the holiday. I don’t have a high opinion of that behavior - look at the CAC adjustments. At least we’re returning to the mean across most metrics.

Now we can get into “Ecomm Summer” properly and prepare for Black Friday.

This week, I am excited to present what might be the best strategic guide Northbeam has ever written, brought to you by a genius on our Media Strategy team.

This is the first in a multi-part series - subscribe for the following issues.

The undisputed, number one problem of all performance marketers: creative volume. 

Given that growth marketers cannot run the same ads to the same potential customers forever, we have a problem. Ads grow stale, and we must constantly make new ones. Everybody in charge of an ad account needs a predictable, scalable source of good-quality “creative” to use in ads.

The problem is getting that creative. If you are reading this, you are probably struggling right now with sourcing new ads. Expensive shoots that don’t perform, blisteringly expensive influencer deals, AI-generated slop that gives you quantity but no quality. 

Most marketers will tell you that user-generated content (UGC) outperforms most other strategies, especially for generating awareness. But, those same marketers will tell you sourcing UGC creative is far more difficult than simple static or graphic assets. 

At Northbeam, we’ve studied thousands of high-performing ad accounts up close. What we know is this: creative velocity is the top determining factor of success for an ad account. 

Northbeam’s top performing brands launch anywhere from 200 to 300 ads a week. Most advertisers in our data set ship around 90. 

How are these marketing teams – some numbering fewer than five employees – launching this much creative? And how, like in the case of Comfrt, are they turning this creative into nine figures of revenue?

The answer: crowdsourcing UGC. 

Our Head of Client Development John Max Bolling has worked with some of the world’s fastest-growing brands for years. What he’s learned about UGC, content creation, and creator management has been distilled into this document.

This is a step-by-step guide on how you can source and test high volumes of UGC ad creative. We’ll cover systems, tools, processes, and even recommend specific hires. You’ll learn which tools to use, how to set them up, and how to automate your workflows. We will cover important pitfalls to avoid and cut straight to master techniques to speed up your execution. 

This is a long and technical document with step-by-step instructions. Bookmark this page. Come back to it. Make checklists as you build each piece, and reference this document as many time as you need. If you have access to our Media Strategy team, they can assist you as well. 

By the end of this article, you’ll know how to build a system for generating hundreds of high-quality UGC ad creatives per week, regardless of your budget. Let’s get into it. 

As the recent trend of “clipping” demonstrates, there is a close relationship between organic and paid ad algorithms on these social media platforms. This crowdsourced UGC process will give you a ton of free organic reach, which will signal to the algorithm that a community is forming. Since the rise of the “For You” style of algorithmic content discovery, most ad platforms will introduce new users to different “communities” via the FYP. This means not only are you getting great ad content, but you are seeding an organic community of content that will be exposed to new audiences. It’s free reach. 

Once this process is live, you will have a near-infinite supply of content that is constantly on-trend. With this, you can test new channels, new angles, new thumbstoppers and hooks, far more rapidly. This can serve your organic social, your television ads, even your product development: studying how these creators behave and use your product can create a feedback loop that’s vital to your business. 

Having an unlimited supply of infinitely creative content removes all barriers from testing new channels, which will open up new audiences and therefore new customers. New customer acquisition is your north star.

The authenticity of user-generated content makes it a naturally high-performing asset. The key is doing it en masse: what we call “authenticity at scale.” By crowdsourcing UGC, you are effectively building a community that supports the strongest marketing form that has ever existed: word of mouth. 

But, could we not do this with AI? There is a key reason why AI influencers do not work at scale: trust. 

Most of your customers don't really like AI, on a political, personal, or philosophical level. Plus legal frameworks to prevent AI influencers have already appeared in places like New York. But on a theoretical level, we aren’t just mimicking influencers. With this process you will build a grassroots network of real humans who feel engaged with and empowered by your brand. That is litigation and future-proof in a way AI influencers are not. 

Crowdsourcing UGC will create a pack of loyal creators and customers who trust your brand. That trust is a powerful fuel for scale, one that cannot be created with AI. 

In this strategy, we will adhere to the AIDA Model, the oldest and most proven concept of customer behavior. Essentially, customers must become aware of your product first, and over time, they become interested and eventually purchase it. 

For crowdsourced UGC to succeed, it must be measured properly. In-platform reporting on your ads (that is, the data you may read directly within Meta Ads Manager) mistracks roughly 50% of your ads per week. Because these ad platforms are “walled gardens” their ability to measure the impact of your ads is limited to what happens on their platform. This UGC process is omnichannel in nature: you cannot track it with in-platform data alone. 

If you are only using in-platform data and you do not have an attribution solution like Northbeam, you are likely to scale the wrong ads and cut budget from ones that are actually driving scale. 

We can see this in the behavior of our top-performing cohort of ad accounts on Northbeam. Across this cohort, 10% of their budget goes to freshly-launched ads per week. Of those ads, 17% of the ads that are actually high-performers are labeled as underperformers according to in-platform data. As much as 29% of the ads that are actually underperforming are labeled as over-performing in-platform. 

On a technical level, the “walled gardens” of ad platforms are best at tracking the final clicks that led to a conversion. These ad platforms cannot easily measure ads that capture attention or generate new customer demand. They can easily track the last click before purchase, however. 

This means in-platform data naturally biases against lower-funnel ads, while ignoring ads that do a better job of creating demand in the first place. If you rely on in-platform data, you end up biasing your spend toward lower-funnel campaigns, which means you pull budget away from the awareness campaigns that feed those lower-funnel campaigns in the first place. It’s a self-destructive spiral. 

Furthermore, Northbeam provides Clicks + Deterministic Views, which allows you to track not just clicks but the views on your ads that eventually led to conversions. This is a massive boon for tracking your awareness campaigns, which often drive view engagement but not clicks. This is common on TikTok and YouTube ads, for example. 

When you layer deterministic view-through attribution on top of clicks-only models, you can attribute as much as 56% more conversions to your ads, most of it top-of-funnel creative. 

This means you have a signal that you can optimize your awareness campaigns against. Tracking views with the same rigor as clicks is the secret sauce of this crowdsourced UGC process.

For a step by step process on how to run your ads, we’ve developed The Daily Ad Optimization Framework, a sister process to this one that you can use daily to make your ad accounts hyper-efficient. 

Now that we’ve established our thesis, the rest of this guide will be more tactical. There are four aspects of the process that must work together.

First, the seeding process: you are constantly recruiting hundreds of small, new creators from TikTok, Instagram, and elsewhere. Nobody is famous. The volume of creators matters more than the pedigree of any one creator. This is a perpetual process. 

You are paying these creators. You pay per video, plus commissions on sales, plus bonuses for hitting a certain number of videos per month. The structure rewards posting, not perfection. 

Take everything your creators make and load it into every ad channel you have, whether that's Meta, TikTok or YouTube. Your creative bottleneck is solved. You have effectively infinite ad angles and no need for elaborate creative strategy, because these creators are guessing and shooting in their own styles.

Aggressively scale spend on the winners. You continue paying creators commission on ad-driven sales so they stay aligned and keep producing. That last piece is the differentiating factor that turns creators into effectively part-time contract creatives for your brand.

The objective: have a constant pipeline of content creators coming in the door, so you have a constantly new crop of content, audiences, and styles to choose from. 

Most people treat creator seeding like an influencer program, which is wrong. They pick a small batch of good creators, vet them carefully, and pay high rates per piece of content. That strategy is closer to a celebrity-style endorsement: you're effectively paying to reach that creator's audience.

We know from experience that the larger the creator and their audience, the less relevant that audience usually is, and the less likely they are to engage or purchase off a product mention. Creators like the influencer model because it gives them higher value deals with less work. 

It's a better deal for them, but it's expensive and it doesn't serve a paid ad strategy. You can target your ideal customers far more effectively than most of these creators can. 

The seeding method is closer to recruiting for a sales floor at a retail store. You need to bring in hundreds of applicants per week, quickly filter for your ideal type of creator, and put them through a low-friction onboarding pipeline. This makes it easy to find creators and tell them what type of content they need to deliver. 

You can be selective later, once you have performance data and understand which creators are making content that outperforms. Then you can build one-on-one relationships and ask for more specific content. A useful way to think about the beginning: you're not hiring 100 creators. You're enrolling 500, to find the 100 who post, to find the 20 whose content actually performs.

There are four sourcing channels to build. 

Run Meta and TikTok ads targeting people who want to make money posting content, and send them to a creator application form. The math works because each new affiliate is both a content unit and a potential customer. They may buy the product, or at minimum they're interested in having it on hand to make ads with.

If you're set up on TikTok Shop, the affiliate marketplace is a captive audience of creators already looking for products to promote. You can mass-invite via the affiliate center, set commission rates, and the platform handles attribution natively. This is easily the lowest-friction channel for getting started.

Anyone who has already bought, posted about you, or tagged you is a pre-qualified candidate. They've shown product affinity, which means their content will read as authentic, and authenticity is the name of the game here. Tools like Archive, which runs around $400 to $2,000 a month depending on plan, automatically capture every tagged and untagged mention across Instagram and TikTok, including Stories. Saral is the Shopify-native equivalent for outbound: it identifies past Shopify customers and lets you send them seeding offers directly by email.

On TikTok you can search creators by niche and follower bands. Target 500 to 50,000 followers at most. That's the sweet spot for hungry, available, relatable creators who are still trying to grow, with good engagement rates and recent posting activity. Modash, at $199 a month and up, is the best discovery tool for creators of this size. Insense also has a marketplace of 75,000 vetted creators. For brute force at scale, you can run virtual assistants on Upwork using a tool like CreatorIQ to scrape and outreach.

The funnel looks like this: 

A creator comes in from one of those sources: a paid ad, a TikTok Shop invite, cold outreach, or a customer email. They land on a page and fill out an application form to join your creator program (bonus if it has a cool name, like Guayaki’s Ambassador program.) An auto-screening automation checks whether they have a TikTok account, and whether they've posted in the last 30 days. Everyone who passes the bar gets auto-accepted. Onboarding is a welcome email, a Discord invite, a free product shipped, and a personalized discount code if you're running gifting. Then, activation: they need to post within 14 days or they get removed from the program.

The key insight: don't pre-vet on quality. The algorithm is a better judge of quality than you are, and you need volume in the pipeline. 

Vet on basic legitimacy instead. Is it a real account? Do they post content? Do they speak the language of your target market? Accept everybody and let post performance do the filtering. The cream will rise to the top in ways that will surprise you. 

Don't pretend you know what the algorithm wants, because honestly, nobody does.

Here's what "small, new, nobody famous" translates to operationally. 

Follower counts between 500 and 50,000: below 500 they may have low engagement and limited reach, and above 50,000 they start asking for flat fees and expecting agency-style treatment. 

Account age under two years is preferred, since newer creators are hungrier and more flexible on terms. 

Posting frequency: they should have been active in the last 14 days, because a dormant account is a dead lead. 

Engagement rate of 3% or more is healthy at this size, and higher is fine, but obsess less about this number than posting recency. This strategy is a content play. Your paid ads will get far more reach than anything they post organically, so engagement rate is mostly a proxy for whether their content is any good. They should also already be posting in formats adjacent to yours: talking heads, get-ready-with-me, demos, unboxings, lifestyle. You're paying for someone who already knows how to point a camera at themselves. 

“On niche”: they don't need to be a perfect match for your business. Pull in creators across tons of verticals. It doesn't matter what they look like, how they talk, or what their primary content type is, as long as there's some level of overlap with your product.

Worrying about your “brand standards” will kill this process before it even starts. Let the creators create - remove your ego. 

The hardest part isn’t sourcing creatives: it’s preventing this program from rotting. The solution is to have a clear, focused process that is run by a creator operations team. 

Most brands can seed 50 creators and get eight of them to post once.

But suddenly, the brand can’t figure out who should be earning what. Payments fall behind schedule. Creators ghost, and the whole program dies within 90 days. I’ve personally experienced this as an operator. 

The unsexy tasks keep this program functioning. Do you have a clean Airtable, HubSpot, or other CRM with every creator's status tracked? Automated payments going out, not manual Venmo? A Discord or other communication channel that gets daily attention from someone on your team? A weekly drumbeat of new briefs, contests, and bonuses to keep posting volume up? 

The Discord in particular is critical because it gives you mobility and speed. When you need something new made fast, being able to speak to creators directly, immediately, is the whole game.

Thus, someone needs to own this crowdsourced UGC process as their primary job. Every creator program, no matter the scale, has to do five things well: onboarding, briefing, tracking, payouts, and retention. 

To properly incentivize creators, the “creator operations” team must master these five jobs:

The single biggest predictor of long-term creator value is whether they posted in their first two weeks. In our experience, if they don't post by day 14, they probably never will. The entire onboarding flow should be engineered to compress that timeline. Here's the minimum viable onboarding sequence. 

Day zero is acceptance: they receive a welcome email with a link to join the Discord, a tracking link and promo code already attached to their account, and a one-page starter brief. Free product ships the same day if you're gifting. 

Day one: an auto-DM in Discord welcomes them by name and points them to a start-here channel with everything they could need: guides, and examples of the video content you're looking for if you have them.

Days three to five: the product arrives, which triggers an automated message: "Got your package? Here are three ideas for your first video." 

Day seven, if they haven't posted: a soft nudge email goes out with examples from top-performing creators in the program. 

Day 14, a final nudge: if they haven't posted by day 21, mark them inactive in your CRM. Don't kick them out. Leave the door open.

You will want to build all of these functions so that someone on your team understands and can troubleshoot them. 

Email automation you probably already have. You’ll need Zapier or Make to wire the application form you build to Airtable, then to an email sequence, to a Shopify order, and a Discord bot like MEE6, Carl-bot, or something custom for auto-welcomes and role assignment. 

This whole sequence should run without anyone touching it. If your team is manually welcoming each creator, you're capped at maybe 50 creators before it falls apart. 

Build everything mentioned in the minimum viable onboarding sequence, and connect everything together so it runs automatically. 

Once they are onboarded, creators will need briefed on what to create. You want them to give a consistent message, but you want them to innovate and create things that will succeed. 

This is the hardest balance to strike, and Comfrt and Ridge both figured it out with the same principle: tight constraints on the message, total freedom on the execution.

A brief should specify four things. 

What the brief should not specify: the exact script, the shot list, the music, the transitions, or their on-camera personality. Creators have to feel like themselves or the content reads as inauthentic, and the algorithm punishes it. You are not controlling for brand identity here. 

What you're doing is carving your own corner into the algorithm, and you need to let creators do that, because they're better at it than you are.

Keep your briefs in a single Notion page or Google Doc that gets updated weekly and pinned in Discord. Drop new briefs on a regular cadence, Mondays work well, so creators always have something fresh to riff off. The simplest brief template worth stealing is a five-line doc, one page, that's it:

Stay tuned for part 2 next week!

📺 Webinar: channel diversification in Q4. Join our friend Andrew Foxwell while he demystifies adding new channels to your mix for BFCM.

📊 ChatGPT owns 92% of AI referral traffic per a 6.77M session study. If you're optimizing for AI visibility, there's really one platform that matters.

🤔 What challenges are you facing as a marketer in Q4? Take this survey from our friends at Foxwell Digital and be entered to win $1,000.

📚 The Operators Podcast has built a sick content portal. Access all their content and get yourself even closer to greatness.

Love this newsletter? Please forward it to your friends, coworkers, or Harry Kane. Subscribe here to get this data in your inbox every week.