Foundations
Part of England's marketing attribution market has no total: supplier presence, method fit and 2027 rework
What five demand signals for attribution can and cannot show
Review five evidence-led signals for marketing attribution demand in England, with transparent scope, limitations and follow-up tests for each one.
These five signals justify research questions, not a claim that England organisations have bought attribution software. A buyer interview, signed brief or paid contract is stronger demand evidence than any item below.
Method: desk review of official, regulator and exact supplier records that state a population or an observable change. Research date: 6 September 2026. Scope: organisations operating or contracting in England, with UK evidence labelled. Exclusions: search volumes, unsourced market reports, job-post counts and global forecasts. Conflicts: no supplier funded this article and no product was accessed or tested.
No audited England purchasing data was found for any of the five signals. The gap is recorded here so the limits of the evidence are clear.
Each signal below is therefore framed as a question, not a result.
What to take away
- Official business population counts differ, so suppliers must define which organisations qualify before claiming reach.
- The CMA study closed in 2020, so it supports a measurement problem but not a 2027 sales opportunity.
- Platform attribution model changes create reconciliation work but say nothing about England purchasing.
- Until buying events are audited, no signal should be described as market demand or likely revenue.
1. Several official business populations can be confused
The ONS 2025 business release counts enterprises and local units from a March 2025 register snapshot. The DBT regional business estimate includes estimated unregistered private-sector businesses and uses head-office geography.
The two headline totals differ by millions of organisations, and that gap is set by the unregistered estimate and the geography rule, not by attribution need. Their difference creates a research need: a supplier must say which organisations qualify before claiming reach or penetration. It does not show how many need or buy attribution.
A supplier quoting either headline owes the reader the definition behind it.
Follow-up test: obtain a buyer frame whose coverage matches the offer and record eligible organisations, responses and verified purchases.
2. Advertising measurement remains a documented problem
The CMA's online platforms and digital advertising study closed in July 2020, so it is not current demand data. Its measurement appendix distinguishes verification, attribution and advertising effectiveness in a UK market investigation. That dated official record supports the existence of a measurement problem, not a 2027 sales opportunity.
The study remains the fullest official UK examination of these markets to date.
Follow-up test: ask recent advertisers which specific decision could not be made and what evidence they commissioned as a result.
3. Platform rules alter credited results
Google's current Ads attribution-model page says model selection changes how conversion credit is assigned and notes that several old rule-based choices are no longer supported. This supplier-specific change can create reconciliation and governance work. It says nothing about other platforms or England purchasing.
Follow-up test: inspect versioned reports from an authorised buyer account and document whether a method change affected an actual decision.
4. Tracking controls constrain observable paths
The ICO finalised guidance on storage and access technologies in April 2026. It covers technologies used in measurement, including tags and pixels, and makes purpose and configuration material to the analysis. That is a signal for privacy-aware measurement design, not proof that a software purchase is required.
Configuration choices, not the software category, decide what the data can support.
Follow-up test: map the proposed events and technologies, then have a named UK privacy and PECR professional assess the configured use.
5. Causal questions need more than monitoring
HM Treasury's QPIE guidance explains, in a government-evaluation context, that observing change differs from attributing it to an intervention and that a strong counterfactual strengthens causal inference. Marketing teams asking an incremental question face the same methodological distinction, although the guide is not commercial market evidence.
Follow-up test: record how often eligible buyers commission a controlled design, abandon the question or accept a weaker descriptive answer.
The next research step is to turn each signal into an eligibility question and seek recent, auditable buying events.
Before you act
- Obtain a buyer frame matching the offer's coverage.
- Record eligible organisations, responses and verified purchases.
- Ask recent advertisers which decision could not be made.
- Inspect versioned reports from an authorised buyer account.
- Have a UK privacy and PECR professional assess configured use.
- Record how often buyers commission controlled designs.
Common questions
What is the difference between the ONS and DBT business population figures?
The ONS release counts enterprises and local units from a March 2025 register snapshot. The DBT estimate adds unregistered private-sector businesses and uses head-office geography, so the two totals cover different populations.
Why is the CMA study not current demand data?
It closed in July 2020. Its measurement appendix separates verification, attribution and advertising effectiveness, which confirms a measurement problem rather than a 2027 sales opportunity.
What does the ICO guidance on storage and access technologies cover?
It covers measurement technologies such as tags and pixels, finalised in April 2026. Purpose and configuration shape the analysis, so it signals a design question rather than a software purchase.
