Diagnose a retention problem

To: Maya Lind From: Staff PM Re: What is happening to retention, and what we should do about it

Bottom line

Retention did not collapse in March, and education is not in freefall. The "fifth" drop in active accounts comes from a metric redefinition. The real leak is in mid-market, where our May price increase and seat audit launched together and revenue has stalled since. I recommend we don't staff two squads on an AI assistant yet. Put one squad on mid-market pricing and seats and one on SMB activation.

1. March was a metric change, not churn

On 2 March the dashboard changed "active account" from "logged in within 30 days" to "created or edited a page within 30 days." Active accounts fell from 2,697 in February to 2,176 in March (−19%), which is the drop you saw. Everything that measures money or contracts says nothing happened:

• Paying accounts rose in every segment in March (total 3,208 → 3,246).
• Total MRR rose 2.6% in March ($573k → $588k), and every segment grew.
• Churned MRR in March was in line with January and February in every segment.

The new definition simply counts fewer accounts as active. The active-to-paying ratio stepped down once and then held: education 86% → 69%, SMB 82% → 65%, mid-market 90% → 73%. The January–February numbers need restating on the new definition. Until then, no one should compare across March.

2. Education is seasonal, with one modest renewal bump

Education "active" fell to 293 in June and 201 in July, then recovered to 296 in August and 461 in September (69% of paying, against 68–70% in March–May). That's a school-year pattern. We have no prior-year data to confirm it, so I'm inferring it from the academic calendar. Education paying accounts and MRR never fell off a cliff.

The one real event is August, the renewal month. Churned MRR was $3,925 against a normal ~$910, so about $3,000 MRR (~$36k ARR) of excess churn. September MRR ($72.2k) is already above July's. Exit reasons are end of academic year (12 of 20), budget cuts (6), and Google Workspace (2). None cite Notion.

One thing to reconcile: education paying accounts rose in August even as churned MRR quadrupled. Either the churn was seat reductions rather than lost accounts, or the account count lags.

3. Notion AI: no evidence yet

Notion launched its assistant on 3 August, five months after the "collapse," so it can't explain March. Since then:

• SMB and mid-market churned MRR in August and September is flat against the prior months.
• Only 5 of 79 exit responses (6%) say "Switched to Notion": 4 SMB, 1 mid-market, no education.

Two months of data is early, and the survey is undated, so I'd watch this rather than rule it out. But it doesn't support a two-squad bet today.

4. The real problem: mid-market, since 15 May

Two changes shipped on 15 May. The plan price rose from $12 to $15 per seat (+25%), and admins got a seat-audit page with one-click removal of seats inactive for 60 days.

May's net figure is inflated by a $31.6k expansion spike, which is the price increase hitting bills. After that, contraction ran 5–9× normal for five straight months. Churned MRR did not move (~$2.7k a month throughout), and mid-market paying accounts kept growing (435 → 460). So we're not losing logos. We're losing seats within accounts that stay. That is my inference from the contraction pattern, and the account-level data should confirm it.

Mid-market grew about $10.3k MRR a month before May. It has been flat since. Against that trend, September MRR should have been ~$360.6k and was $321.9k:

• Gap: ~$38.6k MRR, or ~$0.46M ARR (about 6% of our $640k MRR).
• It grows by roughly $10k MRR every month the pattern continues.

The exit survey agrees: 11 of 15 mid-market cancellations cite "too expensive per seat after the price change." The survey only covers cancelled accounts, so it misses the seat removals, which are the larger leak.

I suspect customers were paying for shelfware. The price rise made them look, and the audit tool made trimming a single click. Some of that contraction is a correction we'd have faced eventually. But we shipped both changes on the same day, so we can't separate the two effects today.

Two open questions: (a) the price increase booked only ~$23k of excess expansion, far less than 25% of a ~$309k base, so I want to confirm whether all accounts have been repriced or some are grandfathered; and (b) how much of the contraction is seats with zero activity versus active seats cut for cost.

5. Second leak: SMB activation

SMB loses the most revenue in absolute terms. Churned MRR runs ~$7.4k a month, about 3.2% of segment MRR monthly, or roughly $1.07M ARR a year. It's masked because new business (~$9–10k a month) covers it. Exit reasons (44 responses):

• Didn't get the team to adopt it: 17 (39%)
• Company closed or downsized: 14 (32%, not addressable)
• Too expensive: 9 (20%)
• Switched to Notion: 4 (9%)

SMB also has the lowest active ratio (65%). Applying the survey mix to churned dollars (a rough assumption from a small sample), adoption failures cost about $34k MRR, or ~$0.4M ARR, a year. That's roughly the same size as the mid-market gap, and it has nothing to do with AI.

Size of the problem

Recommendations

1. Fix the metric this week. Restate January–February on the new definition, show active rate (active ÷ paying) alongside the count, and stop reporting the March figure as churn. Owner: Data/Analytics.

2. Put one squad on mid-market pricing and seats. In two weeks, pull account-level data to split contraction into zero-activity seats versus active seats, by account size. Then run a controlled test across renewal cohorts. Options include: - a lower price (such as $13.50), or grandfathering existing accounts for 12 months in exchange for annual commitments; - a seat-band discount for larger accounts; - a "reactivate seat" flow in the audit page.

I'd keep the audit. Hiding unused seats would cost us trust, and the fix is to make the value of each seat clearer. I wouldn't revert the price blindly until we know whether the lost seats were active.

1. Put one squad on SMB activation. Focus on first-30-day adoption: invite flows, starter templates, and admin nudges when fewer than a set share of seats are active. A target of cutting adoption churn by a third is worth ~$0.13M ARR a year, and we'd see leading indicators in weeks.

1. Education: no product response. Customer Success should confirm the August cancellations and offer budget-constrained schools a renewal option. Also reconcile the account counts.

1. Hold the AI assistant decision until November. Add cancellation date, competitor, and free-text fields to the exit survey. If Notion-attributed churn or win/loss data shows a meaningful rise after August, we revisit with real evidence. A small discovery effort is fine, but not two squads.

To: Maya Lind, CEO From: Staff PM Subject: Retention diagnosis: fix mid-market seat losses and SMB retention before committing two squads to AI

Recommendation

I would not commit two squads to an AI assistant on this evidence. March’s apparent collapse coincides with a dashboard-definition change, education’s summer activity decline largely reverses, and the clearest new revenue problem begins in mid-market after May’s pricing and seat-audit changes—not after Notion’s August AI launch.

There are real losses: persistent SMB churn, a sharp increase in mid-market seat contraction, and an August education renewal spike. Prioritize those, while running a bounded investigation into AI-related competitive losses.

What is—and is not—happening

1. March’s activity cliff is not evidence of a retention collapse

Total active accounts fell from 2,697 in February to 2,176 in March: −19.3%, matching your observation. But on March 2, “active” changed from any login in 30 days to creating or editing a page in 30 days. Those counts are not comparable.

Financial data does not show a corresponding March shock:

• Total MRR rose from $573,323 to $588,463, up 2.6%.
• Paying accounts rose from 3,208 to 3,246.
• Churn and contraction remained close to January–February levels in all three segments.

Under the new definition, total active accounts increased 6.6% from March to September. SMB’s active share of paying accounts was approximately 65% in both months, and mid-market’s approximately 73%.

This does not prove customer retention is healthy: acquisition can conceal customer losses, and account activity can conceal seat reductions. It does mean we cannot infer a March retention collapse—or failure to recover—from this dashboard. We need comparable activity history and cohort retention.

2. Education is seasonal, with a real but bounded renewal issue

Education activity fell from 428 in March to 201 in July, then recovered to 461 in September. Its active share recovered from 31% in July to 69% in September, essentially March’s level. That pattern is consistent with the academic calendar, not sustained freefall.

Financially, education MRR increased from $61,000 entering January to $72,211 exiting September. Nevertheless, August churned MRR rose to $3,925, versus roughly $900 per month otherwise. August gross revenue retention was 94.1%; September returned to 98.3%.

The timing fits annual renewals. Of 20 education cancellation-survey responses, 12 cite the academic year ending, six budget cuts, and two Google Workspace. These are actual cancellations, not merely inactive accounts. We need the August renewal-eligible revenue denominator and prior-year comparisons before judging whether this renewal season was unusually bad.

3. Mid-market’s problem is seat contraction, beginning after May

Mid-market contraction averaged $2,397 MRR/month in January–April. It then rose to:

Account churn remained approximately $2,600–$2,800 MRR/month. We are predominantly losing revenue within surviving accounts, which growing account counts obscure.

Two changes launched together on May 15: a 25% per-seat price increase and an admin tool identifying unused seats with one-click removal. May’s $31,585 expansion spike is consistent with repricing; it is not evidence of stronger adoption. Subsequent seat removal could reflect price sensitivity, newly visible unused seats, or both.

Cancellation feedback supports investigating price: 11 of 15 mid-market respondents cite the price change. But these surveys exclude customers who only reduced seats, and cannot separate pricing from the audit’s effect.

4. SMB has substantial ongoing losses—not a new March or August cliff

SMB loses approximately $7,100–$7,700 MRR/month to cancellations, plus $1,600–$1,700 to contraction. Monthly gross revenue retention is roughly 96%, and net revenue retention approximately 97.3%–97.9%. New sales keep total MRR growing despite losses from existing customers.

Among 44 cancellation respondents, 17 cite failed team adoption, 14 closure/downsizing, nine price, and four Notion. Adoption is a meaningful product opportunity; business closure is less directly addressable.

5. “Notion AI is eating us” remains unproven

Only five of 79 survey respondents cite Notion; none explicitly cite AI. These are unweighted, cancellation-only responses with no dates, not a representative estimate of lost revenue.

Notion’s AI launch was August 3. It cannot explain March’s dashboard discontinuity or June’s mid-market contraction spike. There is no clear August break in SMB losses. Competitive risk may be real, but these exports do not establish it.

Size of the revenue problem

Across January–September, gross churn plus contraction removed $204,460 of MRR movements:

These are summed recurring-revenue reductions—not cash revenue lost during the period, and not wholly recoverable.

The most actionable exposures are:

• Mid-market deterioration: June–September contraction totaled $64,563, about $55,000 above four months at the January–April dollar baseline. That is roughly $660,000 in annualized run-rate equivalent, not a causal estimate or recovery forecast. After expansion, existing-account MRR declined $40,585 over those four months; new sales offset much of it.
• Persistent SMB leakage: expansion offset only part of gross losses, leaving $52,837 in net existing-account MRR reductions across nine months—about $634,000 annualized equivalent.
• Education renewal spike: August churn was approximately $3,000 MRR above the usual monthly level, about $36,000 annualized equivalent, subject to renewal-cohort analysis.

Total company MRR still grew 17.9% from January’s opening balance to September’s close. Growth is masking retention losses.

What to do next quarter

1. Squad one: mid-market revenue retention. With Finance and CS, decompose revenue changes into price, seats, cancellations, and reactivations; compare billing dates and audit usage. Interview downsizers, not just cancellers. Test targeted renewal concessions, right-sized packages, and clearer paid-seat value. Preserve customers’ ability to remove unused seats. Measure cohort seat retention and gross/net revenue retention.

1. Squad two: SMB team adoption. Test onboarding and collaboration workflows that get multiple teammates creating and editing early. Measure activation-to-paid-retention and retained MRR against a control group—not logins alone.

1. CS-led education renewal program. Review August losses by contract value and renewal eligibility; pursue budget-sensitive saves and reactivation before the next academic cycle.

1. Repair measurement; validate AI separately. Backfill comparable activity definitions, publish logo and revenue-retention cohorts, and date/revenue-weight cancellation and downsizing feedback. Run focused competitive interviews and a small AI prototype before making a two-squad commitment.

To: Maya Lind Subject: March retention diagnosis and where revenue is actually at risk

Bottom line: The March dashboard cliff is real as a reported activity change, but it is not evidence that customer retention collapsed. The metric changed on March 2, and paying accounts and MRR continued to grow. There are real, more specific risks: SMB adoption, mid-market seat contraction after pricing and seat-audit changes, and education’s seasonal renewal cycle. The data does not support assigning two squads to an AI assistant on the premise that Notion AI caused a broad retention collapse.

What is—and isn’t—happening

Active accounts fell from 2,697 in February to 2,176 in March, down 19.3%. But on March 2, “active” changed from anyone logged in within 30 days to anyone who created or edited a page within 30 days. That stricter definition plausibly explains the abrupt drop: activity fell by roughly 18–20% in all three segments at once.

The more comparable evidence is what happened after the change. From March to September, active accounts rose from 2,176 to 2,320 (+6.6%): SMB rose 6%, mid-market 7%, and education 8%. SMB and mid-market activity rates were broadly stable over that period. These figures do not show an ongoing collapse—though we cannot make a clean February-to-March retention comparison without recalculating historical activity on one definition.

Education’s summer decline is pronounced: active accounts fell from 428 in March to 201 in July, then recovered to 461 in September. That pattern fits the academic calendar and renewal timing better than a continuous decline. Paying education accounts rose every month, from 625 in March to 669 in September. The activity metric is not the same as paid-account retention, so this is not proof that every education account is secure; it is evidence against calling education “in freefall.”

Revenue also contradicts the broad-collapse diagnosis. Ending MRR rose from $558.6K in January to $640.3K in September, up $81.7K (14.6%). Monthly churned MRR was generally about $10.6K–$11.4K from January through July, rose to $14.1K in August, then returned to $11.1K in September. There is no sustained March deterioration in these figures.

Where the real risks are

• Mid-market pricing and seat contraction: The mid-market plan price rose 25%, from $12 to $15 per seat, and a one-click seat-removal audit launched in May. Mid-market contraction MRR then reached $10.2K in May, $22.4K in June, $16.5K in July, and $12.5K in August, versus roughly $2.4K per month in the surrounding months. That is about $52K more contraction MRR over May–August than the surrounding run rate would suggest. It is a meaningful warning, but not $52K of confirmed churn: contraction may reflect customers removing idle seats, and the report does not isolate price effects from seat cleanup. Mid-market ending MRR was still $12.8K above April’s by September, after the price change and other movements.

• SMB adoption: In the exit survey, 17 of 44 SMB respondents cited failure to get their team to adopt Tandem—the largest single SMB reason. Another 14 cited company closure or downsizing, nine cited price, and four said they switched to Notion. This points to an adoption problem as well as some external business distress.

• Education renewals: Twelve of 20 education respondents said they would review at the end of the academic year; six cited budget cuts. Education churned MRR spiked to $3.9K in August, versus roughly $0.9K in most other months, then returned to $0.9K in September. Prepare for the renewal window rather than treating the summer activity trough as evidence of a year-round collapse.

The exit survey is a sample of 79 cancellations, not a representative churn-rate study. Only five respondents said they switched to Notion (four SMB, one mid-market), and the survey does not say they left because of AI. Notion’s August AI launch is too recent to explain the March dashboard break, and these data do not establish that it is driving the current losses.

Recommended actions

1. Fix the measurement first. Backfill the new activity definition, annotate the dashboard break, and report paid-account retention, logo churn, revenue churn, and usage separately by segment and renewal cohort.
2. Investigate mid-market changes account by account. Separate price increases, seat removals, and cancellations; quantify retained seats and renewal outcomes. Test targeted pricing or seat flexibility where value is at risk before considering a broad rollback.
3. Improve SMB activation. Focus on onboarding, team rollout, and early adoption signals; distinguish preventable adoption failures from business closures.
4. Run an education renewal plan ahead of the academic-year cycle, with early outreach and budget-sensitive renewal options.
5. Do not staff two AI squads on this diagnosis. If AI is strategically important, run a bounded customer discovery and prototype test, with evidence of demand and retention impact as a gate for further investment.