Integrated technical, creative and design products for early-stage and scaling growth teams.
Two quarters with a fast-growing UK wellness brand, and what we aim for with you.
The aims are targets, not guarantees. Where you start decides how far we can go.
Meta’s retrieval and ranking architecture is a Behemoth: an optimised model-space consisting of several multi-trillion parameter models across retrieval and ranking, supported by custom-made inference chips. These systems understand user behaviour better than any team of marketing experts ever will and yet they remain painfully unsatiated.
A system that could in theory serve every single person in the world with the solutions to their most painful problems remains painfully lacking real ammunition, as most advertisers remain fixated on janky custom audiences, shady landing pages and click-bait creative. The task ahead now is more difficult than ever.
Rintrah Systems is a full-funnel growth engineering studio that specialises in building integrated growth systems around Meta. Our fundamental belief is that crafting the Platonic ideal of a full-funnel marketing strategy is now not only technically possible, it is becoming fundamentally necessary.
From creative strategy and landing-page design to ad-account architecture and signal engineering, we audit the customer journey across the whole funnel.
We work inside your existing growth operation to deliver rigorous strategic directive, bespoke AI systems and real long-term results.
The work gets better when the knowledge stays. We learn your economics, remember why decisions were made and carry each result into the next round of work.
One shared backlog. A team that stays accountable. Systems and judgement that become part of the business.
Growth problems do not respect the boundaries between creative, media, conversion, data and operations. We connect these six areas to find the constraint and build the next useful move inside the operation you already have.
Meta can, in theory, optimise towards anything. Your job is to work out the optimisation problem of what that signal needs to be.
Rintrah works with your business to define your strategic needs and then builds a bespoke Meta CAPI, server-side GTM and custom conversion event system.
From how your customers engage with your creatives in-platform, to how they navigate your landing pages, the modern advertising algorithm determines winning ads based on all customer behaviour.
Rintrah maps and re-engineers your entire funnel, from post comment strategy to designing and building custom landing pages.
Ads are a team sport. Creatives can’t be evaluated individually any more, so the goal is a creative portfolio that sits on the Pareto frontier.
Rintrah builds the creative strategy and testing matrix that keeps your pipeline full, gives you a clear testing roadmap and cadence, and finds the local and global maxima across your creative landscape.
Every business is solving a different coordination problem inside the auction.
Rintrah works out the right account structure, bidding strategy, budget scheduling and customer signal journey for your business, so Meta has the data and the freedom to find your customers.
Platform ROAS is Goodhart’s law in action. Attribution can’t separate correlation from causation, so the goal is to estimate the true causal effect of every pound you spend.
Rintrah designs and runs holdout and geo-lift experiments, Meta lift tests and tools like Triple Whale, so you know what your marketing actually caused.
Repetitive coordination should not consume the team’s judgement. We build workflow software around the way the operation already works, including Los Agent and The Forge.
Meta can, in theory, optimise towards anything. Your job is to work out the optimisation problem of what that signal needs to be.
Rintrah works with your business to define your strategic needs and then builds a bespoke Meta CAPI, server-side GTM and custom conversion event system.
From how your customers engage with your creatives in-platform, to how they navigate your landing pages, the modern advertising algorithm determines winning ads based on all customer behaviour.
Rintrah maps and re-engineers your entire funnel, from post comment strategy to designing and building custom landing pages.
Ads are a team sport. Creatives can’t be evaluated individually any more, so the goal is a creative portfolio that sits on the Pareto frontier.
Rintrah builds the creative strategy and testing matrix that keeps your pipeline full, gives you a clear testing roadmap and cadence, and finds the local and global maxima across your creative landscape.
Every business is solving a different coordination problem inside the auction.
Rintrah works out the right account structure, bidding strategy, budget scheduling and customer signal journey for your business, so Meta has the data and the freedom to find your customers.
Platform ROAS is Goodhart’s law in action. Attribution can’t separate correlation from causation, so the goal is to estimate the true causal effect of every pound you spend.
Rintrah designs and runs holdout and geo-lift experiments, Meta lift tests and tools like Triple Whale, so you know what your marketing actually caused.
Repetitive coordination should not consume the team’s judgement. We build workflow software around the way the operation already works, including Los Agent and The Forge.
In William Blake’s mythology, the four Zoas are the great powers of the human mind. Our Zoas are a growing constellation of software and AI systems built to work inside the planetary machinery of modern advertising, from estimating what your spend caused to launching your next round of ads.
In-house causal marketing mix modelling and incrementality testing. We estimate advertising’s contribution and use it to guide budget decisions, where the data and scale support it.
A custom-built ad launcher and creative operations tool for Meta.
A forward-deployed agent configured around the client’s context, decisions and workflows.
Start with an analysis, commission a focused project or bring us into your team for the long term.
Keep more capital in the business while bringing in a team committed to building it with you. For the right partnership, part of our fee can be exchanged for equity, giving us a stake in your long-term success. We agree the work, responsibilities and terms together.
We can: media buying is a service of its own. Most clients keep day-to-day buying with their team or agency and have us design the strategy, review what they do and help analyse performance.
Yes, we can. But one winning ad isn’t the goal. The goal is a winning team: a portfolio of creatives that work together, so your results never hang on a single ad.
Yes, if you want us to. We audit your setup, then either implement Meta CAPI, server-side GTM and custom conversions ourselves, or write the spec for your developer and check it works.
We do the research and persona and angle mapping, build the testing matrix and give feedback on every round. Your team writes the briefs and runs creators, or Los helps them do it at volume.
Our expertise is Meta. We advise on Google and TikTok at a strategic level.
View-only or analyst access to your ad accounts and analytics. We never need admin rights, unless we're running your media buying.
We keep the studio deliberately small so the people designing the work remain responsible for it. Capacity depends on the depth of each engagement, and we will be direct about fit and availability.
Tell us about your business and what you want to work on. We’ll arrange a time to talk.
Notes on Meta, creative and measurement, for the people who have to sign off the budget.
The average Meta ad account is a Jobian tragedy. Media buyers wrestle with a trillion-parameter algorithm that is beyond their comprehension, and indeed any human’s. Despite the towering presence of the algorithmic entity, they are still convinced they can outplay it and beat it at its own game.
They set up an ad set for each audience, an ad set for each concept, an ad set for each audience’s group chat. They try to day-trade their way to a successful marketing strategy by manually controlling every budget and every optimisation, watching every ad with a scornful eye, ready to flick it out of existence at a moment’s notice.
This made sense once. But the algorithm no longer needs to be told what to do. In fact, it is telling you what to do. You need to close your eyes and let go of the wheel.
The usual argument for consolidation is tidiness. That is the least of it. The main reason it works is that inside one CBO, Meta decides the role each ad plays. Given enough creative, the algorithm builds its own funnel: one ad to find people, another to teach them, a third to answer their objections, a fourth to close. It will do this whether or not anybody approved a funnel.
Say you click an ad for a sleep supplement. You have now told Meta you have intent on sleep. Over the next few days it shows you more about sleep: an explainer, a comparison, a testimonial. You click again, a few times, and still do not buy. So it starts answering your objections. Perhaps you think the supplement is missing an ingredient, and Meta finds you the ad that says it does not need it. Then it shows you the offer, and then the discount, until you convert.
Schematic. The sequence varies by person; the principle does not.
Now look at that sequence the way most reports do, one ad at a time. The explainer has a poor cost per purchase, because it is rarely the last thing anybody sees. The offer looks brilliant, because it usually is. Switch off the explainer to save money and the offer gets worse, because nobody arrives at it warmed up. Meta calls this the breakdown effect1, and it has two consequences. Creatives can no longer be judged individually. And an account needs creative for every stage, because the algorithm will go looking for it2.
The second reason is price. What you pay to reach someone is roughly CPM times frequency, which we call CPMR. If three campaigns want the same supplement buyer, all three chase her. You are, in effect, asking Meta to try three different creatives on one woman to see which she tolerates, and paying for each attempt. Put the three in one CBO and only the strongest ad bids.
Illustrative numbers. CPMR = CPM × frequency: the cost of reaching one person enough times to matter.
There is a limit, and it is set by economics rather than by product. Most DTC brands run a separate scaling CBO for products or countries that behave very differently, and they are right to. Products with wildly different margins, order values and lifetime values do not scale well together. Tell Meta to maximise purchases across a chocolate bar and a sleep-tracking ring and it will cheerfully spend the chocolate budget selling rings, report a lovely number, and leave you to work out why the margin went missing.
Ridge, the wallet company, found this out the expensive way. When they went after travel, they first ran it from a separate ad account and pixel, and only 6% of that spend went on non-travel products. After folding everything into one account, one pixel and one purchase event, the figure rose to 40%3. The travel budget had gone back to selling wallets to people who already wanted wallets.
Source: Ridge’s CMO on the Marketing Operators podcast. Read the excerpt.
In our terms, their brand match rate fell: less and less of what people bought was what the ads were about. That is a decent sign the spend had become less incremental. It was harvesting existing demand and calling it growth.
The way to find the line is to ask whether the customers overlap. Someone who buys one magnesium brand is likely to buy another. She is much less likely to be in the market for a smart mattress. Products whose buyers overlap and whose economics are close belong in one CBO, because there they would only be bidding against each other. Products that share neither get their own.
Illustrative overlap: the share of one product’s buyers who also buy the other. Denser engraving, more overlap.
One ABO for testing, where new concepts get a fixed budget and a fair read. One CBO for scaling, with a single ad set, broad targeting and every winner in it. A second CBO only where the economics genuinely diverge: a different margin structure, a different country, a category your existing customers do not already buy. Then stop touching it. Daily fiddling resets learning, and an algorithm that is constantly being corrected never finds out what it could have done.
Consolidate where, and as much as, possible. Not everywhere.
What does living in an attention economy really mean? Beyond scary-sounding essays in The New Yorker and scrolling too much on the gram, we entered the attention economy once human attention became commoditised.
All modern advertising platforms are built on an auction-based algorithm, where businesses bid and compete to get their creative in front of potential customers. In effect, this means that there is a 24/7 live market that determines the actual price of human attention, the same way we determine the price of gold, corn or NVIDIA. This isn’t figurative or theoretical: our society has successfully put a number on our very own precious cognitive experiences.
This hasn’t always been the case. Before the internet, there was no real market for attention. There were several derivative markets – TV, radio, OOH – but prices in those markets were set by ad inventory and guesstimated viewership numbers. Now the supply of ad inventory is nearly fully elastic, and viewership numbers are accurate not only in people but in seconds.
We think most marketers still haven’t fully grasped what this means. Much like the ancient kings of Europe who used to pay for blue dye with its weight in gold, we treat impressions as exotic wonders of otherworldly value. We pay exuberant fees to enchanting merchants, now called influencers, in the hope of securing some goodwill with the universe. These days, as with blue pigment, we can buy impressions at the corner shop, and yet most marketers are still reluctant to do so. For the smart marketer, this creates an excellent arbitrage opportunity.
The fundamental rule of all arbitrage is to buy low and sell high. Typically you do this by finding a buyer and a seller who disagree on what something is worth, be it because of their geography, their currency or their personal beliefs. One of the most common setups is a liquid market, where there are plenty of buyers and sellers and prices are set efficiently, sitting right next to an illiquid one, where there are only a handful of participants and deals tend to be opaque.
The attention economy has both. Meta is the liquid market: millions of advertisers bidding on impressions 24/7, at a price anyone can see. Influencers are the illiquid one: large in nominal value, but priced ad hoc, unprincipled and aggressively negotiated. Most marketers will get suckered here more than once. But it doesn’t have to be that way. All you have to do to not lose your shirt is sense check your influencer deals against the liquid market.
Really, impressions are the only thing you can sensibly price a creator on anyway. You’re almost never going to get proper conversion data from them, and even when you do, you can’t really gauge anything from it. Whether a creator made sales for another brand comes down to seasonality, creative quality, media buying strategy and thousands of other variables. If you want to ask your influencers to run a small Marketing Mix Model on their previous engagements, be our guest. Views and reach, on the other hand, are public, fairly stable and, again, already have an open market price.
On Meta, health and wellness CPMs usually sit between £10 and £18. So one extremely simple way to price creators is to set a target CPM equal to what you’d pay on Meta, and the fee follows:
F creator fee (£)V average viewsp target CPM (£)
Take a £10 target and a creator who usually gets 80k views: 80,000 ÷ 1,000 × £10 = £800.
Fee = views ÷ 1,000 × target CPM. Hatched: £10–£18, the typical Meta CPM in health and wellness. Marker: 80,000 views at £10, £800.
For organic content, that’s basically the whole method (although you can get more nuanced by finding their sponsored-content views, or by comparing how their impressions are trending over time).
One caveat. Most creators also have a floor, a minimum price for making the content at all, however many views they get. Below a certain audience size the formula stops working, because you’re no longer paying for impressions, you’re paying for production. The price is therefore not globally linear in views but only locally so. If a small creator’s floor is £500 and the formula says £200, you aren’t buying attention, you’re buying a video, so price it against what it would cost you to shoot it in-house.
Use the higher of the production floor and the reach-based fee.
F creator fee (£)V average viewsp target CPM (£)Fmin production floor (£)V* crossover views
At Fmin = £500 and p = £10, the two prices coincide at V* = 50,000 views. Below this point, the floor sets the fee at £500; above it, the fee rises with views according to equation (1).
Paid is messier because there isn’t a perfect analogue from paid impressions to organic impressions. The more spend you put behind a creative on Meta, the more impressions it will receive, meaning that in practice you will actually be paying Meta CPMs.
Luckily, with modern account set-ups, you shouldn’t be the final arbiter of your spend. Your winners, your most important ad creatives, will receive the highest spend allocation and will often have a 15–20% lower CPM than your campaign average. Furthermore, the entire point of influencer content and partnership ads is that they are by far the top-performing format in the industry (Meta reports 19% lower CPAs and 13% higher click-through rates on average), so presumably these creatives end up becoming winners. The goal, then, is to price paid influencer deals so that the total CPM, fee and spend included, comes in below your account average CPM and ideally close to your winner CPM.
So say your average creative runs at a £12 CPM, but your mega winners come in at an £8 CPM. Since your influencer fee is an additional cost on top of your paid spend, you need to set the target below your winner rate, say around £6.
A creator doing 80k average views now earns at most a £480 fee. Now say the clip actually turns out to be a mega winner (and so gets an £8 winner CPM in platform) and you put an additional £2,000 of spend behind it. That is 250,000 impressions all in.
Again, your real CPM is total cost over total impressions: £2,480 ÷ 250,000 × 1,000 = £9.92. That’s above your winner rate but way below your average. Good deal.
Deal: £480 fee and £2,000 spend over 250,000 impressions.
Why price off winners? Several reasons. First of all, presuming you also have some in-house video and static winners, spending a ton of cash on an influencer deal is irrational when you can just make more statics and find winners that way. The more technical reason is that winners are where most of your spend ends up anyway. Unless the media spend behind a winner is an order of magnitude more than the influencer fee, the fee will be a big chunk of your total CPM.
There is also a nice operational reason for doing this: winner pricing provides the most aggressive reasonable target, which smart teams will want to hit.
Sometimes you just can’t get someone for under £5,000 a clip and you’re bullish on them anyway. Fine, that’s the whole art of modern marketing. We fully endorse the Don Draper school of spontaneous divine inspiration.
But if you are going to do this, make sure nobody is going to stop the party early. If your influencer deal is expensive, you need to plan your ad spend ahead of the deal.
Say you close that £5,000 influencer deal, the creative comes back, you are popping the champagne, and then your media buyer goes ahead and puts a flat, nauseating £100 behind it. The creative hits a £10 CPM and gets 10,000 impressions, so your total CPM is £5,100 ÷ 10,000 × 1,000 = £510.
Now, we want to emphasise that there is no single influencer in this world, not even Kim Kardashian or Matty Healy, who can justify a £510 CPM on their creative. In fact, it would be far easier to justify paying Kim that same £510 for three seconds of her time.
But say instead you commit £3,000 and get 300,000 impressions. Now your total CPM is £8,000 ÷ 300,000 × 1,000 = £26.67. Still terrible, but at least it’s on the same planet.
Spend relative to the fee is the useful comparison here. The more media spend sits behind the same fee, the smaller the fee’s share of the total cost. Those are the two scales in the chart below.
How the creator fee changes the price of paid impressions.
Spend-to-fee ratio
Fee share of total cost
CPM including the fee
For the £5,000 fee and £3,000 spend above, k = 0.6 and φ = 62.5%. At a media CPM of £10, the fee brings real CPM to £26.67. The chart shows k on its bottom axis and φ on its top axis.
Spend less than the fee and the fee is most of what you’re paying for. With the media CPM held at £10, you need to put a bit more than the fee itself behind the clip before your real CPM even lands back in the normal £10 to £18 range, which for the £5,000 creator means north of £6,000.
Media CPM £10; creator fee £5,000; paid impressions only. Open marker: £100 spend, £510 real CPM. Filled marker: £3,000 spend, £26.67 real CPM. The £18 threshold requires £6,250 spend.
Not only is this a good pricing model, it also gives you a very strong argument internally: if we want this creator, we also have to budget at least X in Meta spend behind them. That gets you more budget, but it also gets the team aligned before the contract is signed. And if we’re not willing to put the spend behind them, we shouldn’t be doing the deal. If you are bullish, you need to be bullish all the way.
In financial markets, every single trade reprices the asset. Influencer markets are more complicated: it is not the initial trade itself but the flip on the liquid Meta impression market that reprices your future trades.
Say your winner CPM on an influencer creative is £10 and your creator averages 200k views. You set a target of £8 and pay them £1,600. The clip comes back at a £17.00 CPM, isn’t eating much spend, and doesn’t seem to play an important role in the funnel.
Your assumption going forward is that £17.00 is roughly what you’d pay in Meta if you did a deal with them again. So next time, if you want to keep working with them, you need to buy even lower. You go in at a really aggressive £5 target, which is £1,000. If they push back, unlike other brands that negotiate on emotion, you hit back with the economic reality of their creative.
We hope this pricing model is useful when thinking about influencer deals more rigorously. That said, it is only a start.
In a sophisticated creative strategy, you are not just looking at CPMs; you are also looking at CTR, CVR and the overall role the creative plays in the funnel. High-CPM creatives that drive overall campaign performance up don’t need to be met with aggressive renegotiations every time. Creative formats and execution also matter as much as who the influencer is.
That being said, the model above can give your team a pricing strategy that puts you ahead of 90% of your competitors and lays the groundwork for your internal models.
The paid examples allocate the entire creator fee to paid distribution and count paid impressions only. Organic distribution, production and usage rights may have additional value; any separate allocation requires an explicit basis. Organic views are used as a negotiating heuristic, not as a predictor of paid delivery. Organic views and paid impressions need not have equivalent commercial value.
The comparisons assume sufficiently comparable impression value for CPM to be a useful cost benchmark. They do not establish profitability or a risk-free arbitrage. The £6 negotiating target is an assumption, not a quantity derived from an £8 media CPM. The £10–£18 range is used in the examples; a dated, market-specific source for it has not been supplied.
The £8 versus £12 example represents a 33.3% CPM reduction. It is distinct from the preceding 15–20% generalisation, for which supporting account data has not been supplied. A lower CPA or a larger spend allocation does not by itself imply a lower CPM. At k = 10, the fee adds 10% to media CPM and is 9.09% of total cost; whether this is minor depends on the chosen tolerance.
Meta’s March 2026 announcement, footnote 8, reports 19% lower CPA and 13% higher CTR from adding partnership ads to business-as-usual campaigns. The stated experiment ran 14–21 June 2025 across approximately 2,400 advertisers. This campaign-level result does not establish that every creator ad will win, that its CPM will be lower, or that it is universally the best-performing format. Meta states that performance is not guaranteed.
A cheaper impression is cheaper for a reason. Stop celebrating falling CPMs and start asking whose attention you are buying.
When CPMs fall, most media buyers celebrate. They should at least pause. A CPM is the price of attention, and when the price of something falls, the usual explanation is that it has become worth less.
Cheap attention is a thumb mid-scroll. It is three seconds of a reel at one in the morning, a glance between two other glances, somebody who could not tell you afterwards what they saw. You can sell things to that attention. The question is what kind of things.
Meta sells attention in an auction. The price of reaching a particular person at a particular moment is set by how many other advertisers want her then, and how much they are willing to pay. Attention that many brands compete for is expensive. Attention that nobody else wants is cheap.
That makes the CPM a signal, not just a cost. A person who is in the market for something considered, who has money, who reads before buying, is wanted by a lot of advertisers, and reaching her costs more. If your campaign is getting unusually cheap impressions, it is worth asking why nobody else is bidding for them.
If a product only sells when attention is cheap, that says something about the product. Nobody enrols in a two-thousand-pound course from a three-second reel. If you were selling something that genuinely mattered to people, you would not want them to glance at it. You would want them to stop, read, and think, and you would happily pay more for the people who do.
Illustrative. Audience A is optimised for the cheapest purchase; audience B for purchase value. Return is 12-month customer value per £1 of acquisition cost.
There is a common idea that Meta shows people things they might be interested in. That is not quite what it does. It works out the sequence of things a person needs to see for an advertiser to reach its goal, and then shows them that sequence. The goal comes from you.
It also does not have to be a purchase. Meta’s delivery system, now largely the Andromeda and GEM models, will optimise for almost any action it can see often enough: a purchase, a subscription, a sign-up, a booked call, an order over a certain value. Give it enough examples and it will find more people who do that thing, at whatever price the auction sets.
This is where authority comes in. The algorithm has no intent of its own. It does not prefer cheap customers or good ones; it finds the ones you describe. If you tell it to find the cheapest purchase, it will find people whose purchases are cheap to get, and those are rarely the people who come back. The only intent in the system is yours, so it had better be the right one.
Illustrative. Each square is a potential customer; its size is what they are worth over a year. The engraved ones are the people the algorithm goes and gets.
The useful number is not the CPM on its own but the most you can afford to pay for it. For every thousand people you reach, some share will become customers, and each of those customers is worth something. Multiply the two and you have the value of a thousand impressions. Anything below that is a good price, however high it looks on the dashboard.
Break-even CPM = customers per 1,000 impressions × value per customer. Stay left of the dashed line.
Run the numbers for two audiences and the cheap one often loses. A buyer who costs twice as much to reach but orders three times as often, or at twice the basket, is the better trade. Most accounts never find this out, because they judge campaigns on CPM and cost per purchase, and both of those numbers reward cheapness.
Tell the algorithm what a good customer looks like. Pass purchase values back through the Conversions API so it can optimise for value rather than volume. Build conditional events for the orders you actually want: a subscription, a basket over a threshold, a first order from someone new. Then report on what those customers are worth, not on what they cost to reach.
Make creative that asks for attention rather than stealing it. Longer videos, specific claims, real reasons: things that a person in a hurry will skip and a person in the market will stop for. That is not a loss of reach. It is the creative doing the targeting, and it tends to find exactly the people who were worth paying for.
And when the CPM goes up, look at who you are now reaching before you try to push it back down.
Cheap attention is cheap for a reason. Pay for the people who are paying attention.