Original Research · Arc Shift Ventures

Startup programs measure everything except the sale.

We went looking for one number from every major corporate startup program — the rate at which startups become commercial customers of the corporate running it — and exactly one program model publishes it.

Four numbers that frame it

60–80%

of US accelerator programs leave startups worse off than never joining, in the largest study yet run — ~750,000 startups, 329 programs. Selection, not acceleration, explains most of the rest.

Baek & Hegde, NBER Working Paper 35063, April 2026 · Grade A · working paper, not yet peer-reviewed; disclosed sample. Conflict disclosed: one author directs an accelerator the paper places in its top-performing tier

280,000

startups received credits from the largest cloud provider’s startup program — US$6 billion over ten years. Published figures on how many became paying customers of that cloud: zero.

The provider’s own 10-year program retrospective, 2023 · Grade B · vendor-published

$31.2B

raised by the alumni of another tech giant’s accelerator; 109,000 jobs across 1,700+ alumni. Published figures on alumni who became cloud or enterprise customers: zero.

The company’s 2025 accelerator impact report · Grade B · vendor-published

>50%

of startups that pilot with one carmaker’s venture-client unit convert to paid follow-up projects — 280+ pilots, 6,000+ startups assessed. The one program type that publishes conversion is the one whose unit of activity is a purchase order.

The unit’s own figures, via an industry interview · Grade B · self-reported

The finding

Among the major cloud-credit programs, corporate accelerators and startup-partnership teams we surveyed, none publishes a startup-to-commercial-customer conversion rate; the venture-client model does. Activity metrics can be hit by the program team alone; a conversion metric needs someone outside the program to commit budget — so those are the numbers that don’t exist. Show us one credits program or accelerator that publishes conversion with a disclosed denominator, and this page gets rewritten.

The counter-model’s own survey: venture-client units adopt 25% of piloted solutions vs 10% without one — vendor numbers (Grade B), but the only conversion rates published anywhere in this field.

Where it breaks

6%

of 1,500+ corporate–startup engagements reached a signed deal when someone actually tracked them end-to-end — the field’s low anchor.

European Innovation Council headline counts, 2025 · EST · a ratio we computed from 1,500+ engagements and 100+ reported deals, not a rate the EIC publishes

81%

of corporates convert fewer than 1 in 4 pilots — the same unowned handover to a revenue owner that kills the program funnel at stage three.

500 Startups, 2017 · Grade A · via Beyond the Pilot

What works

Give the funnel a bottom: one terminal event a budget owner must sign

If you run a program

  • Define the terminal event before the cohort. A PO, a production contract, a referral-attributed account — one event requiring a budget owner outside your team.
  • Publish your conversion rate, with the denominator. It doesn’t need to be high to be credible — it needs to exist.
  • Pay the startup from day one. A paid pilot forces procurement, legal and a cost centre through the pipes while stakes are low.

If you’re the startup in one

  • Treat credits as cost reduction, never as traction. Credits change your burn, not your business.
  • Ask the program its own conversion number. A pivot to alumni-raise totals answers the question too — just differently.
  • Get introduced to the budget owner, not the innovation team. Insist the intro lands on someone who owns a P&L line your product touches.

Numbers that don’t survive a source check

“60% of corporate accelerators fail within two years”

Attributed to a widely-cited industry research firm

Unverifiable as quoted

The number exists only in the headline of a 2019 research page — no sample, method or definition of “fail” is publicly visible, and no underlying dataset has ever surfaced.

“Corporate accelerators cut a startup’s success rate to 8%”

A venture builder’s internal analysis, via startup press

Quoted far beyond its weight class

One internal analysis of a single 2013 founding cohort (baseline 11%, accelerator 12%, corporate accelerator 8%), sample size and “success” undisclosed, never peer-reviewed.

“96% of our accelerator alumni survive, while over 80% of startups fail”

A global tech company’s accelerator impact report

Vendor stat on a zombie baseline

A heavily selected numerator against an unsourced denominator — the “80% fail” baseline traces to no disclosed study, and selection alone predicts high alumni survival.

“80% of the world’s unicorns run on [our cloud]”

A major cloud provider, citing a market-data firm’s unicorn list

True-ish, but a category error where it’s deployed

A cloud market-share observation deployed inside the provider’s credits-program retrospective as program evidence — it says nothing about credits-to-customer conversion, or even whether those unicorns took credits.

Honest limits: the core claim is an absence claim — falsifiable, not provable; unpublished doesn’t mean unmeasured internally; and the venture-client numbers come from the model’s own salesmen. Grades: A = peer-reviewed study or disclosed-sample primary research · B = industry-published or self-reported data, sample not independently verified · EST = reasoned estimate. This page deliberately describes source categories rather than singling out organisations; every named citation is in the appendix of the research version.

Working on a startup program — or stuck inside one?

I connect startups to corporate innovation teams across eleven APAC markets, so I watch daily where program engagement does and doesn’t become revenue. If you want to know what your funnel’s bottom looks like — or whether a program is worth three months of founder time — reach out.

Book a diagnostic Connect on LinkedIn →