I have a simple rule when someone asks me to get involved in a project: first I ask myself whether it’s a problem I’ve already written about, one I’ve thought about without anyone paying me to. If so, it means I genuinely care about it, not just about the opportunity. Devinoso passed that test straight away.
As of this month, I’m a business advisor at Devinoso, an AI-optimised affiliate network for influencers. I want to explain why I said yes, what excites me and, just as importantly, what has to go right for this to become something valuable. No numbers, no promises. Just how I think about it.
A problem I’ve been following for years
Back in 2018 I wrote about the shift from celebrities to micro-influencers. Last year, about the creator economy, where anyone can build their own audience. And in June I wrote about why influencers need proper tracking. The common thread is the same: money in influencer marketing still moves according to old rules.
In short, there are three problems:
- Influencers get paid per post, with no real measurement. The brand pays a flat fee, receives a screenshot of view counts and hopes for the best.
- Brands can’t attribute sales. They know they sold more during campaign week, but not thanks to whom, which post or which audience.
- Classic tracking is broken. Browsers have been restricting third-party cookies for a while, Google has announced it will phase them out too (even if it has pushed back the deadline), and since this spring, with iOS 14.5, apps have to ask users for permission to track them. A lot of the measurement we relied on until yesterday has turned into guesswork.
Put all three together and you get distrust. The brand thinks it’s overpaying. The influencer thinks they’re underpaid. And quite often, both are half right.
The idea: pay for results
I’m describing the idea and the approach here, not a finished product or technical details. Devinoso is designed as a network where influencers promote products and get paid for the sales they generate. The brand no longer buys a post and a hope; it buys a result. The creator no longer haggles over a flat fee, but earns in line with the real impact they have on their audience.
Affiliate marketing itself isn’t new. Online stores have used it for more than two decades, and I’ve been writing about it since 2009. What’s new is the combination: content creators instead of coupon sites, measurement built for a world after cookies, and a layer of artificial intelligence that makes the matches better than a person with a spreadsheet can.
Why AI, and what kind of AI
The word AI gets slapped on everything these days, so let me be precise. I’m not talking about robots writing posts instead of influencers. I’m talking about machine learning applied to data: models that learn from past campaigns and try to answer very practical questions.
- Which creator fits which product, based on what their audience actually buys, not just on follower count.
- Which campaign is likely to convert and which will only bring likes.
- Where the budget should move during a campaign, not after it’s over.
- Which traffic looks suspicious and needs checking before anything gets paid.
It’s the same kind of technology that has been deciding for years what we see in our feeds or which products an online store recommends to us. At Retargeting we worked with exactly this logic, behaviour-based personalisation, only for online stores. Applying it to the relationship between brands and creators feels like the natural next step. A person can assess ten influencers well. A well-trained model can assess thousands and spot patterns we simply don’t see.
But a model is only as good as the data it learns from. That’s why accurate measurement isn’t a detail. It’s the foundation.
Why equity rather than a fee
Two years ago I explained why I work for equity, not for fees, and I’m doing the same at Devinoso. No hourly rate, no retainer, no fixed fee. I take a stake in the company. If I don’t deliver results, I don’t get paid.
I find it almost poetic that my way of working mirrors the product’s model. The influencer gets paid for the sales they bring in. I get rewarded only if the company grows. Everyone has a reason to look at the same thing: the real result, not the appearance of one.
What I look for in a team
I won’t give you names and bios, because that’s not what this article is about. But I’ll tell you what I generally look for when deciding to get involved with a startup, and what convinced me here:
- People who understand the problem from the inside, not just from slide decks.
- An obsession with clean data and honest measurement, not with metrics that look good in a pitch.
- A willingness to say “we don’t know yet” and to test before promising.
- Patience: a network needs both sides, brands and creators, and neither shows up without the other.
The risks, and what has to go right
It would be dishonest to write only about the bright side. An affiliate network lives or dies by trust, and trust is hard to earn and can be lost in a week. Here’s what has to go right:
- Transparency. Creators need to see clearly how their earnings are calculated. Brands need to see clearly what they’re paying for. An algorithm that decides without explaining anything is a black box, and people don’t trust black boxes.
- Fraud prevention. Wherever you pay for results, someone will try to manufacture results: fake clicks, cancelled orders, bought traffic. Here the AI has to work as hard as it does on matching.
- Fair, on-time payouts. For a small creator, a late payment isn’t an inconvenience; it’s a reason to leave.
- Respect for user privacy. People are tired of being tracked. Measurement has to work within GDPR (the EU’s data protection rules) and the new rules in browsers and phones, not against them.
- Quality, not just quantity. If the network pushes creators to sell anything and everything, their audiences lose trust in them. And then we all lose.
None of this gets solved with a single feature. It gets solved with small decisions, made right, day after day.
Why now
Because the problems are visible to everyone and nobody can pretend they don’t exist anymore. The pandemic pushed commerce online, creators have become a serious sales channel, and the old measurement is wobbling. Moments like this are when new models emerge. I don’t know whether Devinoso will succeed; nobody knows that about an early-stage startup. But I do know the problem is real, the approach is right, and I want to be there to help.
What do you think: if you were a brand, would you pay an influencer only for real sales? And if you were a creator, would you accept that deal?