Services · Marketing intelligence - Every platform says it made you money.
They cannot all be right.
For businesses that sell online and spend across several ad channels: we pull every marketing and revenue source into one place, so you can ask, in plain English, what actually made money, and get an answer measured in collected revenue rather than platform claims.
Free to ask. Real reply. Within a day.
The problem - Five dashboards, each telling their own version of the truth.
If you sell online and spend on ads, you know the shape of this. The store says one thing, the email platform another, and every ad channel swears the sale belongs to it.
- Every platform grades its own homework. Each ad channel reports the conversions it claims credit for. Add up what they all claim and you get more sales than your bank account saw. They cannot all be right, and none of them has any interest in being modest.
- The real answer is split five ways. The store knows the orders. The payment processor knows the money and the refunds. The email platform knows the clicks. The ad platforms know the spend. Nobody holds all of it, so nobody can answer the only question that matters: which of this actually made money?
- So decisions run on gut. Budget gets moved on claimed numbers and instinct. Sometimes gut is right. But you are spending real money every day on several channels, and “probably working” is an expensive place to live.
What we build - One place where every number lands. Then you just ask.
The fix is not another dashboard. It is all of your sources unified in one place, measured against money that actually arrived, and an AI that answers questions about it in plain English.
- Every source, one warehouse. Ad platforms, email, analytics, the store, the payment processor: all of it lands in a data warehouse, joined up and matched against collected revenue with refunds netted off. Not what a platform claims. What arrived.
- Questions in plain English. Which channel paid for itself last month. Which campaign brought buyers rather than clicks. Whether that spend increase did anything at all. Asked in a sentence, answered from your numbers, with the working shown underneath.
- A loop, not a report. A report tells you what happened and expires. This is a loop: the system reads the warehouse, proposes changes on the ad platforms, and what happens next flows straight back in. It cuts what is not working and backs what is, every decision logged, and a person approves anything that matters.
The rule that matters - Reads come from the warehouse. Writes go to the platforms.
One architectural rule separates a system you can trust with your ad account from a clever demo that gets it banned.
- Why it matters. Hammering an ad platform’s API to pull analytics is against its terms of service, and it is the actual reason most “my AI got my ad account banned” stories happen. Your ad account carries years of history the algorithms price. It is not a thing to gamble on a shortcut.
- How we build instead. All analysis runs against the warehouse, where the data has already landed through proper channels. The platform APIs get touched for one thing only: bounded, logged actions. The heavy reading and the writing never happen in the same place.
- Bounded, logged, approved. The system chooses from a fixed set of permitted actions, inside spend limits you set, with a person approving anything significant. Every decision is recorded with the evidence it was made on, so you can always see why.
We are building this for our own business first.
This is how we build everything: on ourselves before anyone else. The marketing intelligence capability is currently being built for ScottishWill, the will-writing business our founder runs, which spends on ads and lives on the same question this page opens with.
So this page describes what we build, not a boxed product with a customer list. When we deploy it for a client, it is built around your stack: your store, your email platform, your ad channels. If you would rather buy something off a shelf, we are the wrong firm, and we would rather tell you that here.
Who it's for - Right for a specific shape of business.
This is not for everyone, and the shape matters more than the industry. If the description below reads like your week, this page was written for you.
- Right: selling online, spending across channels. An online store doing real volume. An email platform doing real work. Spend across two or more ad channels. And no single, trusted answer to which of it actually made money. The more channels, the more the claimed numbers disagree, and the more this pays for itself.
- Right: the owner wants answers, not homework. You do not want another dashboard to learn or a weekly export ritual. You want to ask a question in plain English on a Tuesday morning and trust the answer, because it is measured in collected revenue.
- Wrong: not spending yet. If you are not yet spending meaningfully on ads, this is premature. Start with the AI Operations Day instead: a day inside the business, looking at the commercial data, the handoffs and the quality of the sources before anyone builds a reporting layer on top of them.
Questions worth asking before you buy anything like this.
Including the ones that cost us sales, because that is the point of asking.
What is a data warehouse, in plain terms?
One database where copies of all your business numbers land and get joined up: orders from the store, money from the payment processor, clicks from email, spend from the ad platforms. Once they are in one place, questions that used to mean an afternoon of exporting spreadsheets become one sentence. It runs quietly in the background and your team never has to touch it.
Why not just use the dashboards the ad platforms give us?
Because every platform grades its own homework. Each one reports the conversions it claims credit for, using its own rules, and several platforms will happily claim the same sale. The dashboards are useful for running campaigns. They are not a trustworthy answer to which channel actually made you money, because that answer needs your payment data, and they do not have it.
Is this a product we can buy off the shelf?
No, and we would be lying if we dressed it up as one. It is a capability we design and build per business, currently being built for our own will-writing business first, which is how we build everything we sell. What you get is not a login to our software. It is your stack, unified, with the loop built around how you spend.
Will connecting AI to our ad accounts get us banned?
Done wrong, it genuinely can, and that is why our architecture exists. The stories about banned ad accounts almost always trace back to systems hammering the platform APIs to pull analytics, which is against the platforms’ terms. We never do analytics through the ad APIs. Analysis happens in the warehouse; the platforms are touched only for bounded, logged actions, inside limits you set.
What does it cost?
It is scoped per business, because the honest answer depends on how many sources you have and how far you want the loop to go. The only price we publish anywhere is the AI Operations Day at £750 per day, which is also the sensible front door here: a day spent on your commercial data, the handoffs around it and the quality of the sources underneath. Pilots, builds and multi-site engagements are scoped to the work and priced in writing before anything starts, and the full fee is credited against a qualifying pilot or build within 60 days. If there is not a credible pilot worth testing, the Opportunity Map will say so.
Start with a free enquiry
Not a discovery call. Not a pitch with a calendar link. Five questions about how your business actually runs, and it costs nothing. Ross reads every enquiry himself and replies within one working day with a straight first answer: what looks worth automating, and what doesn’t.
If it looks like we can genuinely help, the next step is the AI Operations Day: one working day inside the business, followed by a written Opportunity Map. It shows what is hurting, what should stay human, and the best one or two jobs to prove first. That part comes later, and only if it makes sense for you.
Free to ask. No obligation. Ross replies personally within one working day.