Start with a verdict.
Need a specific question settled? An experiment, an audit, a plan, a written answer you can take upstairs. Fixed-scope engagements from a former Meta ads data scientist, instead of continuous decision automation.
Settle the question with a data scientist. Let the platform keep score.
Fixed scope · Fixed price · Written answer · Not a retainer · Pay at readout
How it works
- You pay at readout, not at signing. You never pay for a document you haven't seen.
- If after week one the data can't support the question, we stop and you owe nothing.
- Every engagement is pre-registered. If the readout doesn't address every criterion we agreed on, the fee is waived.
- If you subscribe to Aviio afterward, every engagement credits in full toward your first year.
Spend & Measurement Audit
Whether your tracking measures what you think it does, where your platforms disagree and why, and what it's costing you.
What we settle
Which conversions each platform counts, which it counts twice, and which never reach it. Where Meta, Google, TikTok, and LinkedIn disagree on the same outcome, and which one to believe. What the gap is worth in dollars at your current spend.
How it runs
Week one: read-only access to your ad accounts, analytics, and a billing or CRM export of the conversion you're judged on, and one 45-minute call on how you define a conversion. Week two: the analysis. Week three: the written report and a readout call.
What you get
A written report, each finding graded settled, likely, or needs a test, with a dollar figure wherever one is defensible. Written to be forwarded upstairs as-is.
Growth Plan
Channel mix, budget levels, and projected outcomes at each level for the next quarter or year. Every projection lands in Aviio and gets graded against what actually happens.
What we settle
Where the next dollar goes across channels, how much to spend at each level, and what each level should return, for the next quarter or the next year.
How it runs
Week one: read-only access to your ad accounts and one 45-minute call on goals and constraints. Weeks two and three: the model, built on your own account history. Week four: the written plan and a readout call.
What you get
A written plan: channel mix, budget levels, and the projected outcome at each level. Every projection lands in Aviio and is graded against what actually happens, so the plan is scored, not filed.
Experiment Design
What to test, how much spend it needs, how long until you know, pre-registered before launch.
What we settle
Which test is worth running, the spend it needs to reach an answer, and how long until you have one.
How it runs
The design is written before launch: the hypothesis, the spend, the duration, and the read that settles it, pre-registered so the answer can't be re-cut afterwards. Creative and audience tests are graded by the platform. Incrementality and lift tests, where the question is whether the spend works at all, are designed and graded with you. The test itself runs on your spend and your calendar, typically two to four weeks for a creative or audience test and six to eight for a geo-holdout. The readout follows within a week of the test closing.
What you get
A pre-registered design and a graded readout. In-platform creative and audience tests from $3,500. Geo-holdout lift studies from $15,000.
LTV Model
What a customer is worth by segment and acquisition channel, how long payback takes, and the CAC ceiling each segment can bear.
What we settle
What a customer is worth by segment and by the channel that acquired them, how long payback takes, and the most you can pay to acquire each segment without losing money on it.
How it runs
Week one: access to your revenue or subscription data and ad accounts, and one 45-minute call on how you define a customer. Weeks two and three: the model. Week four: the written readout and a call.
What you get
Payback curves by segment and channel, and a CAC ceiling for each. Your payback curves become the limits the platform's recommendations respect.
Agency Scorecard
Ninety days of your agency's decisions, graded against what the engine would have done and what actually happened. For the person who signs the invoice.
What we settle
Whether the budget, bid, creative, and targeting changes your agency made on Google and Meta over the last ninety days were the right calls, and what each one cost or earned against the alternative.
How it runs
Week one: read-only access to your Google and Meta accounts and the change-history pull. Week two: we reconstruct the ninety days of changes and grade each against what the engine would have done at the time, and against what actually happened afterwards. Week three: the written scorecard and a readout call.
What you get
A written scorecard for the person who signs the invoice: each decision, its grade, and one net figure.
Something else?
Bring the problem. We scope it, tell you what it would take to settle, or tell you it's not something we do. Scoping is free. The price comes with the scope.
What we settle
Whether your question can be answered with the data you have, and what it would take if it can't.
How it runs
A 30-minute call, then a written scope within a week: the question restated as something testable, the data it needs, the method, the time, and the price.
What you get
A straight answer. If it's not something we do, the scope says so and names what would be needed instead.
Bring us a question
Tell us what you're trying to settle. We'll scope it on a 30-minute call.