is an analytics report out of scope?

The number-pulling is the small part; everything around it is the work. A report worth sending requires accurate tracking set up correctly in the first place, data cleaned of noise and bot traffic, an honest interpretation of what moved and why, and usually a written narrative the client can actually act on. Then there's the recurring cost you don't see until you're in it: every month you're back in the data, every month the client has follow-up questions, and every month you're implicitly accountable for results you may not control. 'A few numbers' is the visible tip of a standing service. Charging for it isn't greedy — it's recognizing that an open-ended monthly obligation is categorically different from a one-time deliverable. If it really were trivial, the client could pull the numbers themselves; the value is in the setup, the interpretation, and the reliability of it arriving every month without fail.

It depends on what you were actually hired to do, and it's worth taking the assumption seriously rather than dismissing it. If your engagement was explicitly to improve a metric, a closing measurement against the baseline is arguably part of proving you delivered, and the client's expectation isn't unreasonable. But there's a sharp line between a single proof-of-result snapshot and an ongoing monthly reporting relationship. Demonstrating that the work hit its goal is closeout; continuing to monitor and report indefinitely is a service. So the fair read is usually: yes to a one-time results report if the project was performance-oriented, no to recurring reporting unless it was named. If the contract is silent, treat the launch snapshot as included and the standing cadence as separate, and tighten the next agreement so 'about results' doesn't get stretched into 'forever measured for free.'

Match the structure to how the client actually wants it. A retainer fits when reporting is genuinely recurring and predictable — monthly numbers plus a short read, every month, on a set cadence. It gives the client reliability and gives you steady income, and it's worth a discount versus ad-hoc because the predictability has value to you too. A per-report rate fits when the need is occasional or unpredictable — they want a report after a campaign, then nothing for a while, then another. Per-report keeps things honest when the cadence is lumpy, but it can be more friction each time. Either way, scope what a 'report' contains, because raw numbers, numbers plus analysis, and numbers plus analysis plus a strategy call are three different products at three different prices. Decide which you're selling before you quote it, or the cheapest interpretation becomes the default.

Invest once in making it fast, then hold the scope. Set up the tracking properly and build a reusable template so each month is mostly populating a known structure rather than reinventing the analysis from scratch — the first report is expensive and the tenth should be cheap. Define exactly which metrics you cover and resist the slow expansion into 'oh, can you also add this dimension, and segment by that,' because every added cut is more time the retainer didn't price. Decide upfront whether you're providing data, interpretation, or recommendations, since each layer is more work and more accountability. And cap the back-and-forth: a report plus a short written read, with deeper discussion available as a separate call, keeps a tidy deliverable from sprawling into open-ended consulting. Profitability comes from a fixed, fast, well-bounded report — not from heroics every month.

You're on the hook for reporting honestly, not for guaranteeing the results, and it's worth being clear about that distinction before you ever send a number. A report's job is to tell the truth about performance and, if you're providing interpretation, to explain plausibly why things moved. It is not a promise that the metric will be good, especially for outcomes shaped by factors outside your control — the client's pricing, their market, their other channels. Conflating reporting with results is how freelancers end up quietly blamed for forces they never owned. So if you offer reporting, frame it as visibility and analysis, and be careful about implying you'll also fix whatever the numbers reveal — that's optimization, a separate service with its own scope and price. Reporting that the numbers are down is doing the job; being held responsible for the numbers being down is a different deal you should only take on knowingly.

It needs a light one, mostly to keep 'now and then' from becoming 'all the time.' Informal, unscoped reporting is where the slow creep lives: a quick update becomes a weekly check-in, becomes a standing expectation you never agreed to, all unpaid because it never felt big enough to charge for. You don't need a heavy retainer for genuinely occasional updates, but you do need a shape — a per-report rate, or a defined support window, or simply a clear statement of what 'now and then' means and what it includes. The lightest version is honesty about the boundary: 'Happy to send the occasional snapshot — if it turns into a regular thing, let's set it up as a proper reporting arrangement.' That one sentence costs nothing and saves you from waking up three months later as someone's unpaid analytics department.