Prioritize the opportunities
Score impact, confidence and effort to turn findings into a ranked list.
An audit usually produces more findings than a team can ship in a year, and the ranking is what turns them into decisions. The method below puts every finding in the same format, converts each one into expected money per day of work, then draws a line at what the team can actually do in six months. Everything under the line gets parked in writing.
Throughout this chapter, say a B2B SaaS at 2M ARR gets 1,000 signups a month. 20% activate, 25% of those pay, which gives 50 new paying accounts at an ARPA of 150, or 7,500 of new MRR a month. The numbers are invented to show the method.
Put every finding in the same format
Collect the output of every chapter in one sheet, one action per line, and merge duplicates. A finding with no evidence goes back to the person who wrote it.
| ID | Stage | Finding | Evidence | Action | Metric |
|---|---|---|---|---|---|
| F-07 | Activation | 80% of signups never reach the first result | Funnel by cohort, session reviews | Rebuild onboarding around the first result | Activation rate |
| F-12 | Monetization | Failed payments have no retry sequence | Billing export | Add retries and card update emails | MRR recovered |
Then give each line a type, because not everything belongs in the same race.
| Type | Definition | Treatment |
|---|---|---|
| Foundation | Without it you can’t measure or ship the rest: tracking on blog CTAs and navigation, an owner for the metric | Out of the ranking, done first |
| Fix | Known problem, known solution | Scored, usually quick |
| Bet | Uncertain outcome | Scored, run as an experiment with a stop rule |
Pick one scoring method and keep it
| Method | Formula | Use it when |
|---|---|---|
| ICE | Impact, Confidence, Ease, each scored 1 to 10 | You need a fast first sort and have little data |
| RICE | Reach × Impact × Confidence / Effort | Actions touch very different audience sizes |
| Expected value | Impact in money × Confidence / Effort in person-days | You have modeled the funnel |
I prefer the third as soon as the growth equation exists. A score of 7 for impact means something different to everyone in the room, whereas “expected new MRR per person-day” can be challenged line by line. Whichever you choose, one method applies to the entire list.
Size impact from the equation
Change one variable in the model and read what moves. Express everything in the same unit, here MRR gained or saved per month.
| Action | Calculation | Monthly MRR |
|---|---|---|
| Rebuild onboarding | Activation 20% to 25%: 250 activated × 25% = 62.5 paying, +12.5 × 150 | +1,875 |
| New-customer price | ARPA 150 to 165 on 50 accounts | +750 |
| Comparison pages | 60 extra signups converting at 10%, twice today’s 5%, because intent is higher: 6 × 150 | +900 |
| Referral ask at first result | 3 extra accounts × 150 | +450 |
| Failed-payment retries | Measured from the billing export | +300 |
| Rebuild the resource hub | 4 extra accounts × 150 | +600 |
Each assumption stays visible so anyone can dispute it. Without one, an impact figure is an opinion with decimals.
Anchor confidence and effort
Confidence comes from the evidence, on a fixed scale, so that enthusiasm can’t raise it.
| Evidence | Confidence |
|---|---|
| Measured on this funnel or tested | 80% |
| Consistent internal data plus customer feedback | 60% |
| Interviews only, or a competitor does it | 30% |
| Opinion | 10% |
Effort is in person-days, estimated by the people who will do the work, with review, QA and tracking included. Have each participant score alone first, then discuss only the large gaps. Sales, product and data all need to be in the room, since each one knows a different part of the cost.
| Rank | Action | Monthly MRR | Confidence | Effort (days) | Score |
|---|---|---|---|---|---|
| 1 | New-customer price | 750 | 60% | 5 | 90 |
| 2 | Failed-payment retries | 300 | 80% | 4 | 60 |
| 3 | Rebuild onboarding | 1,875 | 60% | 25 | 45 |
| 3 | Referral ask at first result | 450 | 30% | 3 | 45 |
| 5 | Comparison pages | 900 | 30% | 20 | 13.5 |
| 6 | Rebuild the resource hub | 600 | 10% | 40 | 1.5 |
In a spreadsheet, with impact in C, confidence in D and effort in E:
F2: =ROUND(C2*D2/E2,1)
Sorted view: =SORT(A2:F30,6,FALSE)Test how solid the ranking is by lowering the confidence of the top lines. At 30%, the price change falls to 45 and ties with onboarding, so it remains in the leading group. Even at 80% confidence the hub scores 12 and stays last, which is useful to know when someone defends it in the meeting.
Correct the score with three checks
The score ranks, and judgment decides. Run these three checks before freezing the order.
Bottleneck. An action upstream of the constraint in the funnel pours more people into the same leak. Here activation is the constraint, so comparison pages wait until onboarding has shipped, even if they would score better with a different assumption.
Dependencies. Foundations come first. Without tracked CTAs on the blog and in navigation, you could not read the comparison pages’ results at all.
Readability. Two changes on the same step at the same moment cannot be told apart. Pricing and onboarding both affect paid conversion, so space them out and give the first one time to be read.
Draw the line with real capacity
Use the capacity computed in the stack and organization audit. Say the team has 70 person-days available for this list over six months. Tracking takes 8, the five leading actions take 57, which makes 65. The hub, at 40 days, doesn’t fit, so it is parked along with a written reason and the condition that would bring it back.
This is the moment people resist, and the moment the audit earns its fee. A list of 40 priorities tells the team nothing.
Deliver the action sheet
The sheet opens with the verdict. On an audit I ran for a B2B SaaS, traffic had good bases and conversion was untouched, so both the analysis and the sheet centered on conversion. State what is solid, where the lever is, what you did not dig into, and where someone else is better placed than you to recommend.
| # | Action | Finding | KPI and target | Owner | Days | Start |
|---|---|---|---|---|---|---|
| 0 | Track blog CTAs and navigation, add acquisition analytics | F-02 | All CTAs tagged | Growth | 8 | Month 1 |
| 1 | New-customer price test | F-10 | ARPA 150 to 165, conversions above break-even | Founder | 5 | Month 1 |
| 2 | Failed-payment retries | F-12 | 300 MRR recovered per month | Finance | 4 | Month 1 |
| 3 | Referral ask at first result | F-15 | 3 extra accounts per month | Customer success | 3 | Month 2 |
| 4 | Rebuild onboarding | F-07 | Activation 20% to 25% | Product | 25 | Month 3 |
| 5 | Comparison pages | F-04 | 60 extra signups per month | Content | 20 | Month 4 |
Under it, add the parked list with its reasons, and the list of topics not covered.
Your next step: copy every finding from your earlier chapters into one sheet using the format above, tag each line by type, and take the top of the list into the 6-month roadmap.