How to assess England's marketing automation market with no official revenue figures

Mistake 1: Pricing the entire business register as a sales forecast

Do not turn England's entire business register into a forecast by applying a guessed subscription fee. That calculation treats organisations with no relevant workflow as buyers.

Set the use case and account filters first, then estimate a reachable group. This overstatement is the costliest mistake because it can distort every decision.

The Department for Business and Trade recorded 5.0 million private-sector businesses in England at the start of 2025. Its business population estimates cover registered and estimated unregistered firms, classified by head-office location. Government and not-for-profit organisations are absent from the main private-sector total.

A narrower denominator is the Office for National Statistics count of enterprises registered for VAT or PAYE. Its 2025 activity, size and location dataset can be divided by industry, employment and geography. Record the chosen population and the method for assigning enterprise location, so the estimate has a clear boundary.

Neither official count identifies organisations already buying marketing automation software. Avoid multiplying the population by an invented adoption share. Filter for workflow, channel, permission, usable information and staff capacity. A workflow-based audience makes the inclusion test visible.

For illustration, a hypothetical calculation starts with 6,000 Greater Manchester business-to-business professional services firms employing 20 to 199 people. After applying its stated assumptions, it yields about 90 possible accounts. The usable database share of 60 per cent, reach of 25 per cent and year-one conversion of 10 per cent are assumptions, not observed demand.

Mistake 2: Letting product names set the market boundary

Use functions, not supplier terminology, to decide what enters the estimate. A contact store, a one-off campaign utility and an automated workflow are not equivalent. State the operating test before gathering names; then apply the same boundary to every account and commercial proposition.

A system belongs inside the working definition when rules, events or models select an audience, trigger an action, coordinate a journey or transfer information between marketing and sales systems. This test centres the estimate on automated activity rather than labels, and keeps products with unlike functions from being grouped together.

An online retailer's abandoned-basket messages and a business-to-business lead team routing leads to account owners can both meet the boundary. Their data, integration and purchasing conditions differ, so estimate those workflow groups separately instead of hiding their needs inside one total.

Charity activity needs its own legal check. The Data (Use and Access) Act 2025 introduced a specific electronic-mail rule that may apply to charitable-purpose soft opt-in. Check the Information Commissioner's direct marketing guidance, updated on 28 April 2026, and do not carry that exception across to ordinary commercial activity without qualified review.

Mistake 3: Converting technology readiness into purchase demand

Do not convert digital-adoption percentages into marketing-automation sales. Those surveys can describe operating context, but their questions do not establish purchase intent for this product. Keep readiness evidence separate and use it to shape account checks, rather than treating it as a conversion rate.

The 2026 demand signals include UK business reports about digitised data and AI use; older ONS evidence records cloud and AI adoption. None asks whether an organisation bought marketing automation. These measures can provide context, but they cannot stand in for product-specific purchasing evidence.

The SME Digital Adoption Taskforce report names barriers including products perceived as designed for larger organisations, switching costs, limited confidence and fragmented support. It is UK-wide policy evidence, not a survey of prospective buyers in this market. Treat simpler onboarding, migration help and clearer exit terms as propositions requiring direct tests.

Check the chosen account group itself: whether customer information is structured, staff can explain current hand-offs and repeated work warrants automation. National adoption measures cannot answer those account-level questions. An interviewee expressing interest is useful evidence to investigate, but it is not a paying customer.

Mistake 4: Treating an eligible account as an approved purchase

An eligible organisation is not yet a sale. A purchase can depend on agreement about process, data, technology and budget, alongside people with time to deliver the work. Map the decision roles and verify that necessary approvals and operating capacity exist before treating a filtered account as likely demand.

In a smaller organisation one person may hold several responsibilities, but each question still needs an answer. An administrator might manage imports, integrations and outbound messages. Use a team role map to identify whose input or approval could influence the buying decision.

Business-to-business marketing also has legal distinctions: PECR separates corporate subscribers from sole traders and some partnerships, which it treats like individuals. Using a person's business contact details also engages UK GDPR duties. Calling a list business leads does not settle its channel, subscriber type, source or lawful basis (the UK rules guide sets out the separate legal questions); seek qualified advice on applicable rules.

Mistake 5: Counting the subscription while leaving the work unpriced

Do not treat a platform's recurring fee as the whole commercial burden. Compare the charging unit with setup, maintenance, integration, migration, support and exit responsibilities. Otherwise, an apparent opportunity may rely on unpaid implementation or leave important customer obligations out of the calculation.

Supplier arrangements vary. HubSpot distinguishes marketing contacts from other stored contacts; Microsoft describes tenant access and capacity concepts for Dynamics 365 Customer Insights; Mautic offers an open-source option that places more hosting and maintenance work with an adopter or partner. These examples explain cost questions, not a supplier ranking.

No account was opened for this assessment, and product pages can change after 5 September 2026. Use the examples to frame questions about cost drivers, plan dependencies and ownership, then check current terms using a documented review method.

A supplier may charge for recurring access, database or usage capacity, implementation, managed operations, or support for a self-hosted system. Include hidden costs by stating who handles data work, configuration, deliverability, training and continuing support, and how the customer can leave.

Security and privacy belong in the operating model. ICO controller and processor contract guidance covers documented instructions, confidentiality, security, sub-processors, assistance, data deletion or return, and audit information. It is marked for review after changes to data law, so check its current position and seek specialist advice.

Match security checks to the sensitivity and scale of the work. The National Cyber Security Centre offers a lightweight SaaS assessment; more sensitive workloads call for review against its full cloud security principles. Consider the contact details and behavioural histories involved rather than assuming campaign creation is low risk.

Mistake 6: Hiding assumptions behind a single market total

Make the estimate reproducible rather than presenting assumptions as official figures. Keep each input attached to its source or rationale, and label it as measured, inferred or hypothetical. Show a range when assumptions vary, and revise those inputs openly instead of giving one unsupported total.

Here is the same example as a low, middle and high range. Every input is assumed for illustration, including a £3,000 annual subscription per account:

This is a range for one segment in one city-region, not a figure for England, and no official source supports any of the percentages. Its use is to show how wide the uncertainty is before anyone quotes a total.

For a platform review, the same method shows which suppliers a segment can reach. Group them by charging unit: those that charge by marketing contacts or profiles, such as HubSpot, Mailchimp, ActiveCampaign and Klaviyo; suites bundled with a wider business platform, such as Dynamics 365 Customer Insights; and self-hosted open-source options, such as Mautic. Each group implies different set-up and support costs for the buyer.

The Greater Manchester example remains illustrative even if its arithmetic is correct. Its 6,000 accounts, 60 per cent usable share, 25 per cent reachable share and 10 per cent conversion are model inputs. They are neither a published market count nor evidence of observed purchasing behaviour.

Keep the count date, definitions, exclusions and source notes beside the calculation. Readers can challenge a range only when they can see which inputs vary between cases. Do not carry an assumption forward merely because it appeared in an older spreadsheet or a global market headline.