One system for deciding what deserves attention next — and learning from what happened. Public evidence, connected data and your own context feed the same loop.
ObserveRecord the signals
What runs itData scienceLLMsAgentic pipelines
ObserveWhat is happening, and where is that coming from?
Signals and data sources, recorded rather than glanced at: your storefront, your connected tools, the market around you. Everything downstream cites what it read here, so a recommendation can always be traced back to a source.
DiagnoseWhat does the data actually say?
Raw signals are messy — duplicates, gaps, seasonality, a tracking change nobody logged. This is the data-science half: clean it, normalise it, and separate what moved from what only looks like it moved.
PrioritiseWhich of these is worth your month?
Findings are ranked against the conditions you are in right now and against the knowledge graph — what your business has already tried, what your market is doing, and what tends to move a business shaped like yours.
PlaySo what do we actually do?
One bounded move, selected from a library of Plays labelled by the metric each is known to move. Not a generated suggestion — a known play matched to the finding, so what it is supposed to shift is decided before it starts.
MeasureDid the metric move, and was it us?
The metric the Play was chosen to move, watched against what else changed at the same time. A move that hits its number while quietly breaking something else has answered a different question.
LearnWhat does this business know now?
The result updates the weights: which levers move your numbers, which never have, and what your team can realistically ship. The next ranking is run against a model of your business that is one turn better informed.
The challenge
Your dashboards are not lying to you. Google Ads will tell you what you spent and what came back, to the euro. What none of them hold is the part that decides whether it worked: what would have happened if you had done nothing.
Conversion climbs, so the checkout work takes the credit. Then the season turns out to explain most of it. The quarter reads well and you end it holding a smaller share than you started with.
The change you made is part of what you are measuring. Compare this month against last month and you are checking your own work with your own ruler — which is how a move that did nothing survives three quarters.
A scan dated last Tuesday is a photograph of something still moving. It has to keep running, against a baseline that moves too, and correct itself when the ground shifts under the last answer.
So we start small. A handful of signals, then keep the ones that turn out to explain something and drop the ones that only ever agreed with you. Nobody can hand you the route from outside the business — it gets found by sailing it and correcting, which is why the sixth month is worth more than the first.
What we keep watching
Every recommendation rests on something being true about your market right now — who buyers compare you against, what they search for, where the demand sits. Those move. When they do, the Play that was worth a month stops being worth a week.
What we record
Named before anything is spent, not reconstructed afterwards when everyone already knows how it turned out. Each one is a signal with a source, so it can be read again next week without asking you anything.
What happens when they move
A competitor enters the comparison, search intent shifts, a category cools — the expected impact is recalculated and the roadmap reranks itself. You find out from us, not from a flat quarter three months later.
That is the difference between a recommendation and a subscription: anyone can tell you what to do once. Staying right is the harder half.
The memory
Best practice is written for the average business, and you are not one. Every correction you make, every Play you finish and every result that comes back is kept — so the advice narrows from what usually works to what has worked for you.
What was tried, what it assumed, and what actually happened afterwards.
Which levers move your numbers, which ones never have, and what your team can realistically ship.
What held becomes stronger evidence. What failed takes the next Play with it.
How it gets used
The same reasoning works whether you own the result or advise the person who does. What changes is where you start and what you need to carry into the next conversation.
For businesses
The work that actually grows the business has no deadline, so it is the work that slips. Curiqo keeps it in view and puts the opportunity with the most impact in front of you.
For agencies
Read a prospect or a client, find the angle you can show your working for, and turn the Plays they approve into the queue you bill against.
Start
No account and no connection. We read the business, form a first working assumption, and come back with the one thing worth discussing.