Portfolio Analytics for Accelerators

Accelerators talk about being "data-driven," but most run their programs on gut feel and founder self-reports. A managing director overseeing 20 companies in a 12-week batch asks each founder for an update deck on Friday. Half arrive late. The numbers aren't comparable because each company defines traction differently. By demo day, the accelerator knows which founders present well, not which companies have real market pull. Billing data solves this — real-time traction signals from Stripe that no founder can spin, cherry-pick, or forget to send.

Accelerator metrics are different from VC metrics

VCs evaluate companies at a point in time: during diligence, at board meetings, at the next fundraise. The cadence is quarterly at best. Accelerators operate on a fundamentally different timeline — they're watching companies evolve week by week over 3–6 months, trying to spot inflection points in real time.

The metric set is different too. A VC portfolio company at Series A has meaningful NRR, LTV:CAC, and cohort retention data. An accelerator company three months into existence has none of that. What it does have is billing data — who signed up, who paid, who expanded, who churned — and the velocity of those events tells you more about traction than any pitch deck.

The core accelerator metrics cluster into two periods: during the program (traction velocity) and after graduation (durability). Most accelerators have a passable process for the first and almost nothing for the second.

Why weekly cadence matters more than quarterly depth

A quarterly NRR number is useless for an accelerator company. The company has existed for one quarter. But weekly new customer count, weekly revenue additions, and weekly trial-to-paid conversions reveal trajectory at the resolution that matters during a program.

A company that added 3 paying customers in week 4, 5 in week 6, and 9 in week 8 is on an exponential curve. A company that added 4 in week 4, 3 in week 6, and 3 in week 8 is stalling. Both might report "10 customers" in their Friday deck. The weekly velocity tells you which one is breaking out and which one needs a pivot conversation.

Traction signals during the program

Accelerator program managers distribute mentoring hours, investor introductions, and workshop slots across 10–30 companies. The allocation decision is implicit and constant: who gets 30 minutes of the partner's time this week? Without objective traction signals, that decision defaults to whoever is loudest, most charismatic, or best at asking for help. Billing data makes it empirical.

First revenue velocity — time to first dollar and ramp speed

Time to first dollar from Stripe is the simplest traction signal. A company that charges its first customer in week 2 of the program is in a categorically different position than one still building at week 8. The absolute timing matters less than the relative ranking within the cohort — if 70% of the batch has revenue by week 6 and Company X doesn't, that's a signal worth a conversation.

After first revenue, ramp speed separates the contenders. How many weeks from first customer to tenth customer? From $500 MRR to $2K MRR? These ramp metrics don't exist in standard SaaS analytics because they're only useful at the earliest stage. But for an accelerator, they're the primary traction vocabulary.

Trial-to-paid conversion and expansion signals

Trial-to-paid conversion rate is the earliest product-market fit signal an accelerator can measure. Industry benchmarks for self-serve SaaS sit between 3–8%, but for accelerator companies the absolute number matters less than the trend. A company converting at 2% in week 4 and 5% in week 8 is learning fast. A company converting at 6% flat is coasting.

Expansion signals — customers upgrading plans, adding seats, or increasing usage tiers — are the secondary confirmation of traction. A company whose first 10 customers all stay on the lowest tier has product adoption but not pricing power. A company whose customers upgrade within the first billing cycle has both.

Early churn patterns — what cancellations reveal in the first 90 days

Churn data at the accelerator stage is noisy — a single cancellation in a company with 8 customers produces a 12.5% churn rate. But the pattern of churn matters even when the rate doesn't. A company whose first five customers all churn within 30 days has a product that doesn't deliver on its promise. A company whose customers churn after 60–90 days has an engagement or onboarding problem, not a value proposition problem.

For accelerator managers, the distinction is operational. The first company needs a pivot conversation. The second needs product coaching. Neither is visible in a Friday update deck.

Demo day readiness assessment

Demo day is the highest-stakes moment in an accelerator program. The companies that present to investors need to answer one question convincingly: "do you have traction?" The problem is that traction is self-defined. A company can claim "3x revenue growth" by cherry-picking its best three-week window. Another can claim "40% month-over-month growth" by counting a $200-to-$280 MRR change.

Billing-verified metrics give the accelerator manager an honest assessment of each company's demo day story. The metrics framework is simple: current MRR (billing-verified, not self-reported), MRR growth rate over the full program duration, customer count and growth trajectory, trial-to-paid conversion trend, and any expansion revenue signal.

Companies cluster into three tiers by demo day. The top tier — typically 15–20% of the cohort — has clear revenue traction: $3K+ MRR, accelerating growth, and expanding customers. These are demo day headliners. The middle tier has early revenue but hasn't broken out: $500–$3K MRR, linear growth, modest customer count. These benefit from focused investor introductions rather than stage time. The bottom tier is pre-revenue or flat: they need more time, not more capital.

The tier assessment should be data-driven, not personality-driven. An accelerator whose demo day lineup reflects billing data outperforms one whose lineup reflects presentation skills — because investors fund traction, not charisma, and the accelerator's reputation compounds on the quality of its introductions.

Post-graduation alumni tracking

Here is where nearly every accelerator fails. The program ends. Founders scatter. The weekly update cadence evaporates. Within 6 months, the accelerator has no reliable data on which alumni companies are thriving, which are struggling, and which have quietly shut down. Alumni tracking becomes an annual survey with a 30% response rate and self-reported numbers that can't be verified.

This matters for two reasons. First, follow-on investment decisions. Most accelerators have follow-on funds or pro rata rights. Without current billing data, the decision to exercise those rights is based on the last conversation with the founder — which may have been 3 months ago and optimistically framed. A company that was "growing fast" last quarter might be churning 8% monthly now.

Second, program improvement. An accelerator can't improve its curriculum, mentor matching, or selection criteria without longitudinal data on alumni outcomes. Which program cohorts produced the best 18-month revenue outcomes? Which mentor pairings correlated with higher retention? Which company profiles at admission predicted post-program success? None of these questions are answerable without ongoing billing data.

The fix is structural. If companies connect their billing systems on program day one, the connection persists after graduation. The accelerator continues seeing billing-verified metrics passively — no founder effort, no update decks, no surveys. When a 2024 cohort company crosses $100K MRR in 2026, the accelerator knows it from the data, not from a LinkedIn post.

The cohort comparison view

Accelerators run multiple cohorts per year, each with 10–30 companies. Over three years, that's 60–180 companies. The most valuable analytics capability at that scale isn't per-company monitoring — it's cohort comparison.

How does the Spring 2026 cohort compare to Fall 2025 at the same point in the program? Is the median time-to-first-revenue improving or degrading? Are trial-to-paid conversion rates higher in cohorts that received the revised curriculum? These questions require normalized metrics across cohorts, computed the same way, with time-aligned comparisons.

North Metric's portfolio view handles this naturally. Each cohort is a group of connected Stripe accounts. The metrics are computed identically across all companies using the same definitions and snapshot timing. Comparing Spring 2026 to Fall 2025 is a filter change, not a data project. The accelerator manager sees median MRR, median customer count, and median conversion rate by cohort — and the trend across cohorts reveals whether the program itself is improving.

For accelerators running follow-on funds, the cohort view doubles as a deal-sourcing tool. Sort alumni by MRR growth rate. Filter to companies above $50K MRR. The breakout companies surface from the data before the founder sends a fundraising announcement — giving the accelerator a first-mover advantage on follow-on rounds.