06-reference/research

services positioning canon morgan baker enns

2026-10-30·research-brief·source: deep-research·by Ray Data Co (deep-research synthesis)
positioningservices-firmsspecializationacquisitionscredibility

A Separate Lineage, and Its Author Says So Out Loud: the Services Canon Locates Positioning in What You Refuse, Not in What You Have

The question

Verbatim: "What does the services-specific positioning canon (Philip Morgan, David C. Baker, Blair Enns) prescribe that Dunford's product-shaped framework does not - is it complementary, or a separate lineage that handles the services case Dunford's step 2 breaks on?"

Context: the direct follow-up to [[2026-09-01-dunford-positioning-b2b-services-application-failures]], whose two attempts to reach this canon died on socket errors. This run reached primary text for all three authors.

Re-aimed, deliberately. The backlog entry named two downstream consumers, both now dead: the raydata.co landing-page rewrite (RDCO dormant per the founder's 2026-09-23 call, no clients, no selling services) and the MAC one-pager (founder 2026-09-30, "We don't need MaC at all"). The question itself is still live, so ## Synthesis for RDCO is aimed instead at the two places it actually bites: the acquisition funnel, where three of five lanes are services firms the founder might buy and reposition, and the founder's own credibility positioning, which is an expertise-based services positioning problem with a sample size of one.

What we already know (from the vault)

What the web says

Convergences and contradictions

Synthesis for RDCO

Answer first: separate lineage, not a complementary layer, and the divergence is structural rather than stylistic. Dunford's framework and the services canon disagree about where positioning lives. Dunford locates it in the market's perception of a bounded artifact whose attributes are fixed and checkable, and her output is a statement. The services canon locates it in a constraint the firm imposes on itself, and the outputs are a refusal list (Baker), a narrowing decision plus a multi-year implementation period (Morgan), and a sales posture (Enns). They are not three steps of one process, because they do not share step 2. The honest decision rule: if the thing being positioned can be changed faster than its reputation can spread, use Dunford; if its reputation spreads faster than the thing can change, use the services canon. A software product is the first case. A services firm, and a single human being, are the second. Morgan's "positioned like a chess piece" versus "one client at a time" is that rule stated as mechanism. The one genuine overlap is Morgan's fifth option, service specialization, which is the same move the parent brief reached independently, and it is the only place the two lineages can be run on the same object.

For the acquisition funnel, this surfaces a real gap in our own rubric, and I want to be careful not to overstate it. Three of the five lanes buy services firms: bookkeeping/CPA, home health/revenue-cycle-management, and vertical agency. The 5-point fit test scores agent/automation leverage, recurring contracted revenue, regulatory moat, owner-operator and books quality, and family-orbit edge. None of the five axes scores whether the target is positioned. That matters because Baker's entire commercial practice is the claim that positioning drives the value of exactly these firms, and Punctuation sells valuation, prepare-for-sale, and sell-side search off a "Sellability Index" alongside its positioning advisory. I checked whether our scores already capture this by accident and they do not: the vertical agencies in the 2026-10-01 scan cluster at 6 and the generalist ones ("SEO and performance marketing agency," "Advertising & design agency, Miami - retainer mix unverified, likely project shop") at 4, but that spread is produced by the recurring-revenue axis, not by positioning. The correlation is real and the causal claim is not mine to make. The actionable piece is that Baker's pre-tests are externally runnable on a target from its listing and website, before any letter of intent - can you buy a list of this firm's stated target, could its principal drop and give you 20 insights, would anyone travel for this - whereas Morgan's introspective process cannot be run on a company you do not own. That is a candidate sixth axis, and it is a buy-side thesis as much as a filter: a services firm with sound economics and no position is the specific thing the canon says a new owner can fix, which is a different and better reason to buy than cheap multiples. The counterweight, from Morgan, is that the fix takes 3 months to 3 years and the flinch risk belongs to whoever owns the payroll through it.

For the founder's own credibility positioning, the canon is sharper and one of its findings cuts against the current cert plan. "Agents in production" is, in Baker's taxonomy, a horizontal: one tight practice applied across verticals. His base rate is that 85% of firms want a horizontal and roughly 5% find a viable one, and the specific failure he names is the claim everybody can make. In October 2026, "agents in production" is that claim. The good news is that the credibility targets doc already escaped this without framing it that way: v2 moved the plan to building in public on an agents/data-on-Snowflake side bet, and in Morgan's five-way taxonomy that is a platform specialization, the one horizontal variant that is addressable, because the platform vendor assembles the list for you. The targets doc's own instruments (Tampa user groups, forum replies, Data Superheroes) are that list. So the plan is better positioned than its label.

The caution is the same source. Morgan's fourth commoditization driver is the platform whole-product ecosystem, whose custodian "wants to make you, a 'vendor', interchangeable with any other vendor in order to lower risk for those who haven't adopted the product. Sure, they need a few incredibly talented experts, but they need a lot more competent, commoditized vendors." Certification, training, and best-practice frameworks are named as the machinery. The founder's cert escalator is banked and each additional certification carries a real $500 bonus plus better deal placement, so this is not an argument against certifying. It is an argument that the certs buy addressability and placement and are structurally incapable of buying differentiation, because the platform's purpose in issuing them is to make holders substitutable. The differentiating asset has to be the thing the platform cannot issue, and all three authors name the same one: proprietary output. Baker's "drop and give me 20" and his separate article on proprietary research; Morgan's small-scale research; Enns' write-to-get-smarter. The build-in-public target ("what I built on Snowflake, what broke, what I'd do next") is already the right shape. Enns' test is the one worth adopting verbatim as the bar for each of those six pieces: "If someone else could have written it then why are you bothering?" A post about what broke on a real Snowflake build passes it. A post explaining agents in production does not.

Why this is in the vault

It closes the explicit unresolved item in [[2026-09-01-dunford-positioning-b2b-services-application-failures]] (open follow-up two, both fetches failed) and converts that brief's labelled inference about a missing refusal step into sourced doctrine. Concretely it proposes a sixth, externally-checkable positioning axis for the acquisition-rescan 5-point test, which currently scores services-firm economics and not services-firm position, and it supplies the acceptance bar for the six build-in-public pieces in [[2026-09-29-nov30-credibility-targets]] plus the reason certifications cannot carry the differentiation load there.

Open follow-ups

Related

Sources

Vault:

Web (all fetched or rendered 2026-10-30 unless noted):