●  NORTHWIND  ·  GROWTH TEAM  ·  QUARTERLY REVIEW

Q3 2026 Growth Review

Strongest signup quarter to date — 1,284 new accounts, up 18% on Q2, with activation climbing to 62% after the May onboarding revamp. The flag this quarter is NRR: down a second consecutive quarter to 104.2%, driven by two logo churns and slowing expansion in sales-assisted. Details, cohorts, and the three Q4 asks below.

OwnerMaya Lindqvist · Growth Lead
Compiled withAtlas (agent)
Versionv4
UpdatedOct 9, 2026 · 09:41 EET
PeriodJul 1 – Sep 30, 2026
Signups, Q3
1,284
▲ +18.0% vs Q2 (1,088)
Activation rate, Q3
62.1%
▲ +2.7 pts vs Q2 (59.4%)
MRR, end of Sep
$48.2k
▲ +11.1% vs Q2 exit ($43.4k)
Net revenue retention
104.2%
▼ −2.6 pts vs Q2 exit (106.8%)
Scopes the trend, funnel, and cohort grid below · showing all segments

Signups & MRR by month, Jan–Sep 2026

All segments · hover for exact values · shaded band is Q3
View as table
MonthSignupsMRR (end of month)

Activation funnel, Q3

All segments · each meter shows conversion from the previous stage

Retention by monthly cohort

All segments · % of activated accounts still active N months after signup · hover a cell
† Feb cohort includes ~430 visitors from the pricing-page A/B test that shipped without a plan gate — low-intent traffic that signed up and bounced. Flagged by Priya in review; excluded from the M1 benchmark. Cohorts from May onward reflect the onboarding checklist revamp (shipped May 12) and hold up 4–6 pts better at M1–M3.

What drove Q3

The carrier-integration templates launch (Jul 22) is the single biggest driver. The launch week was our largest signup week ever (168), and September closed +12.5% MoM at 467 signups — templates now appear in 41% of new-account first sessions. Second, the onboarding checklist revamp from May 12 kept compounding: visitors who track a first shipment inside day one convert to paid at 3.4× the base rate, and Q3 activation landed at 62.1%, up 2.7 pts. Third, the annual-plan push on the reworked pricing page lifted activated→paid conversion to 26.8% (from 24.1% in Q2), with annual now 31% of new MRR.

What's off track

NRR is the problem, and it's concentrated in sales-assisted. We lost Meridian Freight ($1.9k MRR) and Baltica Chartering ($1.1k), both citing missing EDI X12 support in their exit calls — the same gap that stalled two open expansion deals. Expansion revenue slowed to $2.1k/mo from $3.3k/mo in Q2, and sales pipeline coverage sits at 2.1× against a 3× target. Self-serve retention is healthy but M1 drop-off remains our steepest cliff: ~33% of activated self-serve accounts go quiet in month one, and the Feb pricing-test cohort (see grid) shows how sensitive that number is to traffic quality.

Q4 asks

  1. One backend engineer-quarter for the EDI X12 connector
    Top-cited churn and stalled-expansion reason. $14.2k ARR directly at risk across 6 accounts; Meridian has said in writing they'd return. Scoped with platform team at ~9 eng-weeks.
    owner: Dana / Platformdecision by: Oct 24
  2. $18k/quarter paid-search budget on the container-tracking-API cluster
    Organic already ranks 4–7 on the head terms; paid test in Aug returned signups at $61 CAC vs $89 blended. Cap at $18k with a kill threshold of $80 CAC at week 6.
    owner: Maya / Growthdecision by: Nov 3
  3. Ratify the activation definition change before annual planning
    Proposal: activated = first shipment tracked within 7 days (currently 14). Tightens the metric by ~5 pts but correlates 0.81 with month-3 retention vs 0.62 today. Needs sign-off from Product and Data so 2027 targets are set on the new definition.
    owner: Priya / Datadecision by: Nov 14

Review notes

Changelog for this document — comments resolved in the doc thread.

v4Oct 9 Recut the funnel and cohort grid by segment after Dana's comment on the activation dip — self-serve vs sales-assisted now split throughout, which shows the dip is a self-serve traffic-mix issue, not product. maya + atlas
v3Oct 8 Added the Feb cohort footnote after Priya flagged the M1 retention outlier; Atlas re-pulled the cohort table from the warehouse with pricing-test traffic tagged. atlas
v2Oct 7 Corrected NRR to exclude annual-prepay true-ups (−1.3 pts vs v1). Q4 asks rewritten with owners and decision dates per leadership template. maya
v1Oct 6 First draft compiled by Atlas from warehouse.growth_daily, Stripe MRR export, and Metabase #412; Maya edited the narrative sections. atlas + maya