By the end of week 1, a good agency has already made you do work. They have requested ad account, analytics, and CRM access. They have asked for 12 months of spend data, your churn numbers, and intros to 2 or 3 customers they can interview. You feel slightly pestered. That is a good sign.
What week 1 usually looks like instead: a pleasant kickoff call, a recycled onboarding deck, and then quiet until the next scheduled check-in. The difference between those two week 1s predicts the whole engagement.
This is the accountability framework I would hand any founder signing with an agency, ours included. Put it in the contract conversation. If you are still deciding whether an agency is the right move at all, read the readiness test for hiring a growth agency vs building in-house first.
What Should a Growth Agency Deliver in the First 90 Days?
A growth agency should deliver three things in 90 days: a written diagnostic with baseline CAC and LTV math by day 14, live experiments with weekly reporting by day 45, and a scaled winner plus a documented channel playbook by day 90. Each phase has a deliverable you can see and a metric that proves it happened.
Here is the full checklist in one table.
| Phase | Deliverables you should see | Metric that proves it |
|---|---|---|
| Days 1 to 14 | Access secured, tracking audit, baseline CAC/LTV doc, ICP and message review | A baseline number you both signed off on |
| Days 15 to 45 | First creative and channel tests live, weekly reporting starts, first kill/scale calls | Number of experiments launched and killed |
| Days 45 to 90 | Winners scaled, incrementality check, written channel playbook | At least 1 channel or angle beating baseline |
| Day 90 | Keep-or-fire review with documented results | Your decision, made on data |
Notice what is not on the list: a brand refresh, a 40-page persona deck, or a promise that “results take 6 months.” Those are common, but they are not deliverables.
What Happens in Days 1 to 14?
Days 1 to 14 are for access, audit, and baseline math. The agency secures every account, audits your tracking and attribution, and produces a one-page baseline: current CAC, LTV or payback period, and conversion rates by funnel stage. They also pressure-test your ICP and messaging against real customer language.
The baseline document matters more than founders expect. Without an agreed starting number, month 3 reporting becomes a negotiation about what “improvement” means. With one, it is arithmetic.
A tracking audit usually finds something. In our work across $50M+ of managed paid media, broken or double-firing conversion events show up in most new accounts we open. Fixing that before test spend is often the cheapest win in the engagement.
The ICP review should involve your customers, not just your opinions. If the agency never asks to hear a customer’s voice, they are planning to advertise to a guess.
What Should Be Live by Day 45?
By day 45, first experiments should be live and reporting should be weekly. That means creative tests running against your control, at least 1 channel test if the diagnostic justified it, and a standing weekly report tied to the baseline metrics. The first kill-or-scale decisions should already have been made.
Creative is usually the first lever because it moves the most. According to Nielsen’s research on advertising effectiveness, creative quality contributes 47% of sales impact, more than reach, targeting, or any other element they measured. An agency that tests audiences for weeks before testing a single new creative angle has the priorities inverted.
Expect the early reads to take 2 weeks or more per test, and be suspicious of anyone calling winners faster on thin data. Meta’s own documentation says an ad set needs about 50 optimization events to exit the learning phase. Budget and patience have to cover that, or every “result” is noise.
The weekly cadence is non-negotiable from here forward. Not a dashboard link. A short written narrative: what ran, what it showed, what changes next week.
Want this timeline as a structured engagement?
Our 90-Day Jumpstart runs this exact arc: diagnostic in the first 30 days, structured experiments from day 31, scaling and a 6 to 12 month roadmap by day 90. It is standalone, with no long-term contract after.
Book a Free Strategy CallWhat Does Days 45 to 90 Look Like?
Days 45 to 90 are for scaling winners, checking that they are real, and writing it all down. Budget moves toward what beat the baseline. The agency runs at least a basic incrementality check, such as a geo holdout or a spend pause, to confirm the channel is driving new revenue rather than claiming credit for it. The phase ends with a written channel playbook.
The playbook is the asset that survives the relationship. It should document what was tested, what won and lost, the audiences and angles that work, and the next 2 quarters of priorities. Our 90-Day Jumpstart closes with a 6 to 12 month roadmap for exactly this reason: you should own the learning whether you continue with the agency or take it in-house.
Then comes day 90, and this is the part founders skip. Put a formal keep-or-fire review on the calendar before you sign anything. Decide in advance what “keep” requires: for most teams, at least 1 channel or creative angle beating baseline with believable attribution, losers killed fast, and reporting you never had to chase.
What Are the Red Flags in Agency Onboarding?
The 3 biggest red flags are a strategy phase still running on day 60, no live experiments by day 30 (or at minimum a finished diagnostic with launch dates locked), and reporting that does not start until month 2. Each one means you are funding process instead of progress.
A few more worth watching:
- The deck is the deliverable. Strategy documents that restate your own onboarding answers back to you, with stock-photo polish.
- Metrics drift toward soft numbers. Reporting that leads with impressions and engagement when you are paying for pipeline.
- No kill decisions. An agency that never kills a test is either not testing or not telling you what failed.
- The team switched after the pitch. Senior people sold it, junior people run it, and your emails now get answered in 3 days.
Vague scope is usually the root cause, and it is set at signing, not at day 60. The questions in our breakdown of growth marketing agency pricing for SaaS (who staffs the account, what ships monthly, what metric reporting centers on) are the same questions that prevent most of these problems.
How Do You Hold an Agency to This Timeline?
Hold the agency to it by putting the phases in writing before you sign. Attach the table above to the agreement, name the day-14 baseline doc and day-45 experiment count as deliverables, and book the day-90 review on day 1. An agency confident in its process will agree without flinching.
Hesitation at that request is itself the answer. Good operators want a defined finish line because they expect to clear it.
We built our entry engagement around this exact structure because founders kept arriving burned by open-ended retainers. If you want a 90-day arc with a diagnostic, live experiments, and a roadmap you keep either way, that is what the Jumpstart is for.
Like this? Get the next one.
Short emails. New posts as they ship.
