The day-90 review: keep, kill or scale

Why the review matters more than the last week of work

The review turns a quarter of activity into decisions you can act on. Without it, you know roughly how the sprint felt but not which parts paid for themselves.

There's also a practical reason. During a sprint, changes get attention every week. Once the sprint ends, the attention goes away, and any change that depends on someone remembering will fade. The review is your one chance to turn the good changes into routines before that happens.

Step 1: compare against the baseline, honestly

Start with one question: did the target figure move?

Put three numbers side by side:

Then classify the result plainly:

• Hit: at or above target.
• Moved: clearly above baseline but below target.
• Flat: within normal weekly noise of the baseline.
• Worse: below baseline.

In the example, the result is "moved": a real gain of five points, short of the target. That's a perfectly good outcome, and calling it a failure because it missed the target would be as unhelpful as calling it a triumph.

Check the guardrail too. If close rate rose but average sale value fell by a similar proportion, the business may have discounted its way to the gain.

Step 2: trace the cause through the decision log

Next, work out why the figure moved. Go through the decision log week by week and match each decision to what happened in the following two to three weeks.

You're looking for patterns like these:

• A clear step change. Close rate jumped two weeks after the new quote template went live. Strong evidence.
• A steady climb. Close rate rose slowly as follow-up compliance climbed from 40% to 85% of quotes. Good evidence for the follow-up routine.
• No visible effect. The new email signature changed nothing measurable. Probably harmless, possibly pointless.

Be cautious with outside factors. A seasonal rush, a competitor closing or a price change elsewhere can move numbers on its own. Note anything like that in the review so you don't credit the sprint with something it didn't do.

Step 3: sort every change into keep, kill or scale

List every change made during the sprint. Then put each one in a column.

Keep changes that worked and cost little to maintain. Examples: a reply-time standard, a quote template, a referral ask at job completion. For each, write a short procedure: what it is, who does it, how often, and how you'll check it's still happening.

Kill changes that didn't move anything, or that cost more time or money than they returned. Examples: an ad campaign with a cost per enquiry above what a sale is worth, a weekly social post that never produced an enquiry. Killing things is as valuable as keeping them; it frees time for the next sprint.

Scale changes that worked and could do more with more effort. Examples: a money page that produced enquiries cheaply, suggesting two or three more pages for related services; a reminder message that brought back 5% of consented past customers, suggesting a second reminder later in the year.

A worked example of the sort, for a lead flow sprint:

Step 4: write the handoff

The review produces a short document, three or four pages at most:

1. The result against baseline and target, in one paragraph.
2. What caused it, citing the decision log.
3. The keep list, each item with a procedure and an owner.
4. The kill list, so nobody quietly restarts those items.
5. The scale list, with a rough estimate of the effort each would take.
6. Accounts and files: confirmation that every account, page, template and list sits in the business's name, with access listed.

The last item matters more than it looks. Whoever helped run the sprint, inside or outside the business, the business must be able to carry on without them.

Step 5: decide what the next quarter is for

The review ends with a choice. There are usually four sensible options:

• Consolidate. Spend the next quarter making the keep list routine, with no new sprint. Good when the team is tired or the gains are fragile.
• Scale. Run a second sprint in the same lane, built on the scale list. Good when one lever clearly works and has room to grow.
• Move on. Run the next sprint on the next-narrowest stage. Good when the first constraint is fixed and another is now the bottleneck. Rerun the bottleneck exercise with fresh numbers; the answer often changes once one stage is widened.
• Stop. If the sprint showed the real issue lies outside marketing and sales, such as capacity, pricing or cash flow, address that first.

Whichever you choose, keep the 13-week scoreboard running, even if only monthly. It's the cheapest insurance against gains fading.

Common mistakes on day 90

• Declaring victory on feel. If the four-week average is within noise of the baseline, the honest verdict is "flat".
• Keeping everything. Each kept change has an ongoing cost in time. Keep what earned its place.
• No owner for kept changes. A procedure without a name attached usually fades within six weeks.
• Starting the next sprint the next Monday. A short pause lets you check that the kept changes hold on their own.
• Losing the files. Make sure templates, lists and logins are stored where the business controls them.

Getting help with the review

Every BusinessBooster.ca sprint ends with this review, written up and handed to the owner, as described on the services page. If you ran a sprint yourself and want an outside view on the result, an hourly session is often enough; see pricing. To book one, call, text or email.