06-reference/research

cpa bookkeeper aging cohort tooling spend

2026-09-01·research-brief·source: deep-research·by Ray Data Co (deep-research synthesis)

The CPA retirement wave is real and measurable; the per-firm software spend it supposedly frees up is not published, and it is the wrong prize anyway

The question

"What's the demographic shape of US CPAs / bookkeepers retiring or aging out 2026-2030, and which segment has the highest per-firm tooling spend that AI augmentation could displace?"

Surfaced 2026-05-02 from the boring-business-to-disrupt thread as the vertical-expansion wedge beyond dbt shops. It now sits under a live lane: the RDCO acquisition funnel already carries a Bookkeeping/CPA lane with two scored candidates.

What we already know (from the vault)

What the web says

Convergences and contradictions

Synthesis for RDCO

The demographic answer is yes, with a caveat about precision. There is no publicly available year-by-year US CPA retirement curve for 2026-2030 — anyone quoting one is interpolating. What exists is a measured slope: retirement went from 15% to 24% of voluntary departures between 2020 and FY2024 in a 1,073-firm benchmark, it concentrates in larger firms, and 64% of surveyed firms are 21+ years old, which is consistent with an owner cohort that founded in their thirties in the early 2000s and is now mid-fifties (INTERPOLATION — the 52-53 average-age figure is secondary). For RDCO's purposes that is sufficient: the wave is real, it is accelerating, and it does not need a decimal place to justify Channel 3 of the outreach plan.

The segment answer is the CAS practice inside a $750K-$5M NCF firm — but the displaceable spend is labor, not software, and this is the finding that should change how the lane is pitched. The CAS survey describes a segment that assembles a bespoke tool stack per client (only 46% integrated, 53% generic configs), does the integration labor by hand (only 42% have a dedicated team), and gives that labor away — 45% of practices charge no separate setup fee and absorb it into a fixed monthly fee. That absorbed labor is the real target. Against it, the software line item is small: even the unverified 4-8%-of-revenue figure is a quarter the size of the 28.7-34.4%-of-NCF non-owner salary line the same firms carry. Displacing tooling spend is a $40K-$80K-a-year prize per firm at the high secondary estimate. Absorbing the labor is a $300K-a-year prize at the $1M NCF band. Any RDCO positioning that leads with "we replace your software" is aiming at the smaller number.

The firms themselves have already named the mechanism, which is unusual and worth exploiting. 39% of MAP respondents with a capacity plan intend to "expand client load without adding staff." The CAS survey supplies the quantified version: practices committed to continuous technology investment serve a median of 100 clients versus 67 and post NCF per professional of $181,440. At the $17,867 average revenue per client, the 67-to-100 gap is roughly $590,000 of additional annual revenue per practice on a fixed professional headcount (ESTIMATE — assumes the marginal client prices at the median, which top-performer data suggests is conservative). That is the sellable number, it is derived from the segment's own benchmark rather than from an RDCO model, and it is a growth story rather than a cost-cutting story — which matters, because 40% of firms admit they cannot currently track efficiency gains from technology, meaning a cost-savings pitch lands on buyers with no instrument to verify it.

For the acquisition lane specifically, this sharpens diligence on the Palm Beach candidate. That practice is 74-83%-tax-services-shaped like its band peers, with two employees plus the owner. The band ratios say roughly a third of NCF is non-owner salary; the agent-absorption thesis is arguing about that third. The 245-client individual tax book and 20-30-client monthly accounting book are the structured production surface. But note the constraint MAP puts on the upside: tax is a seasonal, deadline-bound production line, whereas the 67-to-100-client expansion math comes from CAS, which the sub-$1.5M bands only carry as a major revenue line 8-14% of the time. The margin-expansion story is strongest for a CAS-heavy book and weakest for a tax-heavy one, and the funnel is currently full of tax-heavy books. That is a real qualification on the "agent-driven margin expansion closes the milestone gap" claim carried in the August scoring note, and it should be raised before an LOI, not after.

Why this is in the vault

This calibrates axis 1 of the 5-point fit test in [[2026-08-30-shortlist-refresh-scored]] with industry ratios for the CPA lane, and it changes the diligence question on the Palm Beach candidate from "can agents absorb the work" to "is this book CAS-shaped or tax-shaped," which the scoring note does not currently ask. It also supplies the pre-listing targeting logic for Channel 3 of [[2026-08-30-pre-market-outreach-channel-plan]]: filter the DBPR/FICPA list toward firms carrying a CAS or monthly-accounting book, not toward the largest firms, because retirement pressure concentrates upward but PE absorbs that end first.

Open follow-ups

Related

Sources

Vault

Web (PRIMARY — full text extracted and read)

Web (SECONDARY — attributed but not verified to primary)