nrr-recovery-coach

Coach a B2B SaaS team through diagnosing and fixing broken Net Revenue Retention. NRR below 100% kills enterprise valuation faster than slow growth, because public-market and acquirer math weights NRR as a forward growth proxy. Most companies wake up to the problem 2 quarters late.

This coach runs in two modes: diagnostic (NRR is broken; figure out why) and recovery (we know what's broken; execute the playbook).

When to engage

Trigger when:

• "NRR dropped from 115% to 92%"
• "Gross retention is below 88%"
• "Expansion revenue is flat or shrinking"
• "Downsells are exceeding upsells"
• "Logo churn spiked to 15% annual"
• "Net new ARR is below board plan because of churn"
• "Board wants a recovery plan in 30 days"
• "We're heading into a fundraise / sale and NRR is killing the multiple"
• "CS is overwhelmed; renewals slipping"
• "We don't even know our NRR — different teams report different numbers"

Do not engage for: pure new-logo growth problems (different skill — sales / pipeline coach), product-led growth funnel optimization (different — PLG coach), pricing redesign as a strategy (use saas-pricing-auditor first, then this).

Diagnostic sweep — what's actually wrong?

Step 1: agree on definition

Most "NRR is broken" conversations are actually "we don't agree on what NRR is" conversations. Lock down:

• Numerator: Renewing ARR + Expansion ARR (upsell + cross-sell) − Downsell ARR − Churn ARR.
• Denominator: Starting ARR (cohort at start of period).
• Window: Trailing 12 months (TTM-NRR) is the most defensible board-and-investor metric. Quarterly cohort NRR is useful for trends but noisier.
• Exclude new logos acquired during the window. NRR is about the existing book, not new sales.
• GRR (Gross Retention) = same as NRR but with expansion zeroed out. Reveals pure logo + downsell health.

Get these three numbers separately:

• TTM-NRR (cohort-based, expansion-on)
• TTM-GRR (cohort-based, expansion-off)
• Logo retention rate (just count of customers, not revenue)

If TTM-NRR is dropping but GRR is stable, the problem is expansion (not churn). If GRR is dropping, the problem is retention.

Step 2: decompose the variance

Build a simple waterfall: Starting ARR → +Expansion → −Downsell → −Logo Churn → +Reactivation = Ending ARR.

Track each component as a % of starting ARR over 4 trailing quarters. This isolates the regression:

• Expansion declining quarter over quarter? Sales motion broken or product roadmap stalled.
• Downsell increasing? Customers actively reducing seats / dropping modules; usually pricing or value-perception problem.
• Logo churn rising? Retention problem; map to cohort and segment.

Step 3: cohort the churn

Group customers by signup quarter, segment, ARR band. Look for:

• ICP drift: are recent cohorts churning faster than older ones? Sales is closing wrong-fit deals.
• Onboarding cliff: do customers churn at month 3-6 (TTV failure)?
• Renewal cliff: at month 12 (annual contract — the big-bang churn point)?
• ARR-band concentration: small accounts (\x3C $10K) churning at 30%+ is normal; mid-market (>$25K) churning at 15%+ is broken.
• Industry concentration: if 60% of churn is one vertical, you have a vertical problem (product gap, ICP misfit, or competitive loss).

Step 4: name the dominant root cause

Pick exactly one (or at most two) primary drivers. Spreading the recovery sprint across 5 root causes guarantees no progress.

5 churn-driver categories:

1. Product gap — customers churn because the product no longer (or never did) solve the job. Often surfaces in QBRs or churn surveys.
2. Time-to-value failure — customers churn before they ever get to value (typically 0-6 months in). Onboarding broken.
3. ICP drift — sales closing wrong-fit deals. Usually shows up in cohort-level retention regression.
4. CS under-investment — book has grown faster than CS headcount; coverage broken; QBRs missing; renewals reactive.
5. Competitive displacement — better/cheaper alternative in market; customers leaving for a named competitor.

5 expansion-leak categories:

1. No value-metric pricing — customers can grow usage without paying more (per-org pricing where seats / events / GB would be more aligned).
2. No upsell motion — sales doesn't have an expansion playbook; CSMs aren't compensated on expansion.
3. Missed seat-growth — customer adoption happened but expansion conversation never did (no quarterly review, no usage-trigger alerts).
4. Missed module-attach — additional product modules exist but cross-sell motion absent.
5. Contract-mechanic friction — annual contracts with no mid-term expansion mechanism, or aggressive lockup of usage limits.

The 90-day recovery sprint

Week 1-2: Stabilize and triage

• Build at-risk list: every customer with renewal in next 90 days, scored R/A/G.
• Personally pause exec-level on R-list (CEO + CRO + CSM-leader) — every red account gets an executive sponsor.
• Pause non-urgent product roadmap: the 2 most-cited churn reasons get fast-tracked into roadmap.
• Stop hemorrhaging: if a competitor is winning displacements, get a competitive battle-card live within 7 days.

Week 3-4: Saves and stabilization

• Run save calls on every R-list account: discovery (why considering leaving), package (price concession + roadmap commitment if needed), close (signed renewal extension).
• Track "saves" weekly: count, $-saved, win-rate. Celebrate publicly.
• Acceptable concessions: 5-15% discount, longer payment terms, free additional modules for term, custom support; never give away the platform.
• Forbidden concessions: indefinite price lock, custom code that creates technical debt, bespoke SLAs not aligned with platform.

Week 5-8: Recovery motion build

• CS coverage audit: ratio (CSM : ARR), pod composition, named-account assignment. Right-size to cover all > $X ARR with named CSM.
• Onboarding rebuild (if TTV is the issue): time-to-first-value target (e.g., \x3C 14 days from signature), milestone definition (what does "activated" mean?), automated nudges for stalled rollouts.
• Expansion motion: CSM compensation tied to expansion (typically 10-20% of OTE), upsell discovery questions in every QBR, usage-trigger alerts when customer is approaching plan limit.
• Pricing audit: are downsells happening because customers are over-priced for usage? Run a value-metric analysis (use saas-pricing-auditor).

Week 9-12: ICP correction and reset

• ICP review: which segments are retaining best? Sales should be biased toward those.
• Sales compensation review: if reps are paid only on new logo with no clawback for early churn, fix that. Standard: 12-month churn clawback (reduces commission for accounts churning within 12 months).
• Renewal forecasting: implement leading indicators (NPS, usage health score, executive engagement, support ticket count, login frequency).
• Quarterly NRR / GRR / Logo retention dashboard for the board, with cohort breakdown.

Boardroom narrative

When NRR is broken, the board wants 3 things: diagnosis honesty, specific plan, leading indicators.

The diagnosis statement

Pick the one root cause. Don't list five. Example: "GRR fell from 92% to 85% over the last 3 quarters. Cohort analysis shows the regression is concentrated in customers acquired Q1-Q3 of last year, primarily SMB tier, primarily in [vertical]. Root cause: ICP drift — sales over-rotated to a segment with insufficient TTV given current onboarding."

The plan

3-5 specific actions, each owned by a named exec, each with a measurable outcome and timeline.

Leading indicators

What metrics will move first if the plan is working?

• Save rate on at-risk accounts: target 60-70% within 30 days.
• Time-to-first-value reduction: target -50% within 60 days.
• Expansion pipeline coverage: target 2x of expansion target within 90 days.
• Logo retention by cohort: improvement visible at month 3-6 mark.
• NRR / GRR turn: typically a 2-3 quarter lag from when fixes go in to when reported metric improves.

What not to say

• "We need more salespeople" (when retention is the problem).
• "We need a price increase" (without value-metric analysis).
• "We're investing in product" (without specific feature commitments).
• "Customer success is hiring" (without ratio targets).

When NRR can't be recovered

Sometimes the diagnosis is: ICP is fundamentally wrong, product doesn't solve a sticky problem, market segment is dying, or competitive landscape has decisively shifted. In that case, NRR recovery is the wrong frame — pivot is.

Signals NRR recovery won't work:

• 4+ consecutive quarters of declining GRR with no specific root cause stable across cohorts.
• Multiple segment-cohorts all churning, not concentrated.
• Competitor is structurally cheaper/better and gap is widening.
• Usage data shows core feature engagement falling across the board.
• Customer feedback dominated by "we don't really need this anymore" not "fix X feature."

When you see these signals, escalate: this isn't a recovery sprint, it's a strategic-direction conversation. Use saas-pricing-auditor + saas-acquisition-prep-coach to evaluate options (relaunch, repositioning, sale).

Common mistakes the coach should flag

• Counting new logos in NRR. Inflates the metric, hides the problem.
• Tracking simple NRR (revenue-weighted) without cohort NRR. Hides cohort regressions.
• Defining churn as "explicit cancellation". Customers downgrading to a free tier, going dormant, or moving to month-to-month are churning in slow motion.
• Treating GRR as a vanity metric. GRR is the load-bearing metric; NRR depends on expansion which depends on GRR.
• Hiring CSMs as a fix without process. A new CSM with no playbook is just a more expensive support agent.
• Heavy discounting at renewal as a save tactic. Trains the customer to negotiate every cycle; permanently impairs ARR.
• Skipping post-mortems on churn. Lost opportunity to learn the root cause.
• Comparing NRR to public-comp benchmarks without segment match. Mid-market enterprise SaaS at 110% NRR is normal; SMB SaaS at 90% can be normal; PLG bottom-up 100% is great. Match comps.

Renewal forecasting

A leading indicator system that flags renewals 90 days out. Components:

1. Usage health score: 0-100 based on logins, feature adoption, usage volume vs plan limit. Update daily.
2. Stakeholder map: champion (the buyer / power user), executive sponsor (the budget holder), end users (count + activity). Score on count + activity.
3. Engagement score: QBRs attended, NPS, support tickets, executive touch.
4. Contract risk flags: auto-renewal language, change-of-control, early termination rights, last QBR > 6 months ago.
5. At-risk classification: R = 1+ red flags + low usage + missing champion. A = 1 red flag. G = no flags.

Forecast at 90 / 60 / 30 days; resolve to renewed / extended / lost / downgraded.

Integration with other coaches

• saas-pricing-auditor: if downsell is a major driver, pricing redesign is upstream.
• expansion-revenue-coach: if expansion is the leak, this is the focused playbook.
• customer-onboarding-coach: if TTV is the issue.
• saas-acquisition-prep-coach: if NRR recovery is being driven by an upcoming exit / fundraise.
• b2b-saas-pricing-coach: for value-metric pricing analysis.

This coach takes 90 days to show results; cohort metrics lag 2-3 quarters. Set expectations accordingly.