 

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%)

All Self-serve Sales-assisted 

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 

| Month | Signups | MRR (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