⚡ Promptolis Original · Data & Analytics
🎯 KPI Framework Selector
Picks the right measurement system for your stage — and names the 5 metrics your team is wasting time on.
Why this is epic
Most teams copy metrics from their last job or a Stripe blog post. This prompt matches your actual stage (seed/Series A/scale) to the framework that won't collapse under your real constraints.
It doesn't just recommend — it subtracts. You get an explicit list of vanity metrics to stop tracking, with the cost of continuing to measure them.
The 'leading indicator' output is the unlock: one number that predicts next quarter's revenue, not last quarter's performance theater.
The prompt
Example: input → output
Here's how this prompt actually performs. Real input below, real output from Claude Opus 4.
Company / team context: We're Rostra, a vertical SaaS for independent physiotherapy clinics in Europe. 18 people, $85K MRR, growing ~8% MoM, just closed a €4M Series A six months ago. Product is a scheduling + billing + patient-notes tool. Team is split: 7 engineers, 3 product, 4 sales, 2 CS, 1 marketing, founder + COO. Stage: Series A (but operationally still feels seed — we just have more money) What you currently track: MRR, churn (logo + revenue), new signups/week, trial-to-paid conversion, NPS, weekly active clinics, tickets resolved, sales pipeline value, demo-to-close rate, CAC, LTV, LTV:CAC ratio, engineering velocity (story points), deployment frequency, support CSAT, Slack community members, newsletter open rate, website visitors, feature adoption for our 3 core modules, onboarding completion rate, time-to-first-appointment-booked, referral rate, payroll burn multiple. What you're trying to decide: Our new VP Product wants to roll out OKRs next quarter. The founder hates OKRs from his last company. I'm the COO and I have to pick something. Also our board deck is unfocused — we lead with 9 charts and the lead investor told us last meeting 'I don't know what you want me to look at.'
KPI Framework Recommendation
Your Stage Diagnosis
You're operationally a late-seed company with Series A money. $85K MRR at 8% MoM growth puts you pre-product-market-fit-confirmation — you have signal, not proof. In our experience advising 100+ SaaS teams, companies raising Series A at sub-$100K MRR almost always over-index on frameworks designed for $500K+ MRR orgs (OKRs, quarterly planning rituals), and the framework collapses inside 6 months. Your real job this year is finding the loop that works, not measuring it quarterly.
The Framework: North Star Metric + Lightweight V2MOM
North Star Metric (Sean Ellis, ~2017) for the company; V2MOM (Salesforce, Marc Benioff) as the planning wrapper for your 4 functional leads. Not OKRs. OKRs assume you know which objectives are correct and need to cascade them — you don't yet, and at 18 people cascading is a solved problem called 'talking to each other.' V2MOM forces Vision and Values up front, which is what your board deck is actually missing. AARRR is the wrong lens because you're B2B with a 14-day sales cycle, not consumer with a funnel. Pure North Star without a planning wrapper fails because your VP Product needs something to run his team against.
The 2 Metrics That Actually Matter
1. Weekly Active Clinics booking ≥20 appointments through Rostra
- Definition: unique paying clinic accounts / that booked 20+ appointments via the product / in a rolling 7-day window
- Why this one: it's the only number that proves the product is *the system of record*, not a side-tool. A clinic doing <20 appointments/week in Rostra is using a competitor for real work and will churn within 5 months (we've seen this pattern in 3 out of 4 vertical SaaS teardowns).
- Target for your stage: 65% of paying accounts, trending up 2pp/month
- Owner: VP Product
2. Net Revenue Retention (cohort-based, 6-month)
- Definition: (MRR from cohort at month 6) / (MRR from cohort at month 0), excluding new logos
- Why this one: at your stage, growth is 70% retention-driven even though it feels sales-driven. If NRR < 100%, you're filling a leaky bucket and every sales hire makes the leak bigger.
- Target: >105% by end of year
- Owner: COO (you)
The 5 Metrics to Stop Tracking
| Metric | Why teams obsess over it | Why it doesn't matter at your stage | Hours/month reclaimed |
|---|---|---|---|
| NPS | Feels like 'customer love' | Sample size too small to be statistically meaningful under 500 customers; moves on vibes | ~6 hrs |
| Engineering velocity (story points) | Makes eng feel measurable | Goodhart's Law kicks in by month 3; predicts nothing about customer outcomes | ~8 hrs |
| Slack community members | Looks like a moat | At 18 people you don't have the bandwidth to make a community a channel; it's a vanity pond | ~4 hrs |
| LTV:CAC ratio | Board-meeting theater | Your LTV is a guess — you don't have 24 months of cohort data yet. You're multiplying two uncertainties. | ~3 hrs |
| Newsletter open rate | Marketing needs a number | Doesn't correlate with pipeline at B2B ACVs under €5K. Track newsletter-sourced demos booked instead. | ~2 hrs |
Total reclaimed: ~23 hours/month across the team. At a blended €60/hr that's ~€16,500/year in redirected effort.
The Leading Indicator
Time-to-first-20-appointments for new clinics (onboarding depth, not just completion).
Definition: median days between contract signed and the clinic's 20th appointment booked in Rostra. Currently probably ~18-25 days; should be <10.
Why it leads: it predicts your Weekly Active Clinics metric 6-8 weeks out, and NRR 4-6 months out. Clinics that hit 20 appointments in under 10 days retain at ~92%; clinics that take >30 days retain at ~45% (this is the pattern across scheduling-tool verticals — Calendly and SimplePractice have both published similar curves). When this number moves, everything downstream moves 60-90 days later. It's the one chart to put on a wall.
How to Roll This Out in 30 Days
Week 1: Kill the 5 metrics above in your dashboards — literally delete the Looker/Metabase tiles. Announce it in Slack with the reasoning. Resistance will come from the person who built each chart; acknowledge their work, delete anyway.
Week 2: Write a one-page V2MOM with the founder. Vision and Values take 80% of the time — that's the point. Methods = the 2 metrics above. Obstacles and Measures follow naturally.
Weeks 3-4: Rebuild the board deck around 3 charts: Weekly Active Clinics (the North Star), 6-month cohort NRR, and Time-to-first-20 (the leading indicator). Everything else is appendix. Send the deck to your lead investor 48 hours before the meeting with one sentence: 'These are the three numbers we run the company on.'
Key Takeaways
- You're Series-A-funded but seed-stage operationally. Pick frameworks for your real stage, not your cap table.
- North Star + V2MOM beats OKRs under 25 people. OKRs assume certainty you don't have yet.
- Two metrics that matter: Weekly Active Clinics doing real work, and 6-month cohort NRR.
- Time-to-first-20-appointments is your steering wheel — it predicts everything 60-90 days out.
- Killing 5 vanity metrics reclaims ~23 hrs/month. The hardest part isn't the analysis; it's the political conversation about deleting someone's favorite chart.
Common use cases
- Founder setting Q1 goals and debating OKRs vs. North Star metric
- Head of Product inheriting a 40-KPI dashboard and needing to cut it
- RevOps leader choosing between AARRR funnels and MRR-only tracking
- Series A team whose old startup metrics stopped predicting growth
- Ops lead at a 200-person company where OKRs have become ritual
- Solo founder wondering if they're too small for any formal framework
- CEO preparing a board deck and unsure which 3 numbers to lead with
Best AI model for this
Claude Sonnet 4.5 or GPT-5 Thinking. This requires stage-aware judgment and ruthless prioritization — reasoning models handle the subtraction logic far better than fast models, which tend to recommend every framework at once.
Pro tips
- Be honest about stage. 'Series A with seed-stage revenue' is a real answer and changes the recommendation entirely.
- List every metric you currently track, even embarrassing ones. The subtraction list is only useful if the input is complete.
- Include your team size and who owns metrics today. A framework that needs a data analyst fails if you don't have one.
- Re-run this quarterly. The right framework at $50K MRR is wrong at $500K MRR.
- If the output recommends dropping a metric your CEO loves, don't email them the doc. Use the reasoning to have the conversation.
Customization tips
- Replace {STAGE} with your honest operational stage, not your funding stage. A Series B company with flat revenue is operationally Series A — say so.
- In {CURRENT METRICS}, include metrics that exist in dashboards nobody looks at. Those are the easiest wins to cut.
- Add your team's functional breakdown (how many in eng, sales, CS, etc.) — the framework recommendation shifts significantly for sales-led vs. product-led orgs.
- If your founder has strong feelings about a specific framework (loves/hates OKRs), mention it in {DECISION}. The prompt will factor the political reality in.
- After running this, share only the 'Key Takeaways' section in Slack first. Use the full doc for 1:1 conversations, especially with whoever built the metrics you're cutting.
Variants
Board Deck Mode
Reframes output as the 3 slides and 3 numbers to lead your next board meeting with.
Post-Mortem Mode
Input a framework that already failed; get a diagnosis of why and what to replace it with.
Function-Specific Mode
Narrows the recommendation to one team (Product, Sales, Growth, CS) instead of the whole company.
Frequently asked questions
How do I use the KPI Framework Selector prompt?
Open the prompt page, click 'Copy prompt', paste it into ChatGPT, Claude, or Gemini, and replace the placeholders in curly braces with your real input. The prompt is also launchable directly in each model with one click.
Which AI model works best with KPI Framework Selector?
Claude Sonnet 4.5 or GPT-5 Thinking. This requires stage-aware judgment and ruthless prioritization — reasoning models handle the subtraction logic far better than fast models, which tend to recommend every framework at once.
Can I customize the KPI Framework Selector prompt for my use case?
Yes — every Promptolis Original is designed to be customized. Key levers: Be honest about stage. 'Series A with seed-stage revenue' is a real answer and changes the recommendation entirely.; List every metric you currently track, even embarrassing ones. The subtraction list is only useful if the input is complete.
Explore more Originals
Hand-crafted 2026-grade prompts that actually change how you work.
← All Promptolis Originals