Marketing automation service standards: a review framework

Marketing automation service standards are a useful lens in a marketing automation platform review: assess promises across the full journey, not only headline uptime. Apply this framework to products such as HubSpot, Salesforce Marketing Cloud, Marketo, Klaviyo, Mailchimp, ActiveCampaign, Braze and Customer.io; record any published commitments separately from the organisation's measured end-to-end results. Use service standards to describe measurable operating promises within the organisation's control; each needs a service, period, population, calculation, evidence source, owner, exception rule and response when the threshold is missed.

What to take away

Define the service before the number

Draw the path from source event to audience decision, message, destination and preference update. State which systems and hours are in scope. For each measure, identify the clock start and stop, exclusions, sampling method, data latency and reporting period. Keep a failed measurement visible rather than silently dropping it from the denominator.

The GOV.UK Service Manual tells relevant public-service teams to minimise downtime and plan for failure. An England-based private organisation can borrow its questions about monitoring, capacity and recovery, but should not claim government-standard status.

Choose standards tied to harm

Availability matters only if it covers the function the team needs. Define whether it measures journey entry, audience calculation, message dispatch, preference capture or reporting. Set separate expectations where failure has different consequences.

Data freshness should record the age of the last complete and reconciled transfer rather than only the time a connector ran. Preference handling needs an end-to-end measure from receipt to effective suppression in every relevant destination. The ICO's direct marketing guidance, read alongside the UK GDPR and the Privacy and Electronic Communications Regulations (PECR), supplies the UK regulatory context, while a qualified lawyer must decide the exact obligation and acceptable operating limit.

Use a recovery measure for time from detection to safe containment, then another for tested restoration. Do not combine them: a journey may be stopped quickly but take longer to reconcile. Record duplicate, lost and delayed events during recovery.

Cover security, accessibility and support

Security standards can specify privileged-access review, log availability, critical-alert response and recovery-test completion. The NCSC cloud security principles support structured questions about identity, audit and operational security. They do not prescribe one universal target or prove a supplier configuration is safe.

Accessibility standards should name content and user routes, such as messages, forms, preference pages and error handling. WCAG 2.2 offers testable success criteria. Record manual checks and assistive technologies rather than reporting only an automated score.

Support standards need severity definitions, acknowledgement, ownership, update frequency and escalation. Keep the organisation's internal response separate from a supplier's contractual support promise. For processor arrangements, the UK GDPR and the ICO's processor-contract guidance identify requirements and subjects including assistance, security and end-of-service; commercial counsel must review the broader service terms.

Use an auditable measure card

For each standard, record: metric name, operational risk, population, period, numerator, denominator, exclusions, data source, evidence owner, threshold, alert route and review trigger. For a duration, mark numerator and denominator as not applicable and define the timestamps instead.

Investigate a miss before changing the target. Note whether the cause was demand, data, configuration, supplier dependency, staffing or a bad definition. Repeated exclusions and manual corrections are findings in their own right.

Review standards when the channel, audience, supplier, processing purpose or service hours change. Retire a measure only when its underlying risk disappears or another named measure covers it. The next useful action is to take one existing dashboard number and write its complete measure card; any field the team cannot fill is evidence of a control gap.

Before you act

Common questions

What should a marketing automation service standard include?

It should describe measurable operating promises within the organisation's control. Each standard needs a service, period, population, calculation, evidence source, owner, exception rule and response when the threshold is missed. A supplier uptime claim alone is not a standard for the whole journey.

How should availability be defined for marketing automation?

Availability matters only if it covers the function the team needs. Define whether it measures journey entry, audience calculation, message dispatch, preference capture or reporting. Set separate expectations where failure has different consequences, rather than using a single uptime figure.

What should be done when a standard is missed?

Investigate a miss before changing the target. Note whether the cause was demand, data, configuration, supplier dependency, staffing or a bad definition. Repeated exclusions and manual corrections are findings in their own right. Review standards when the channel, audience, supplier, processing purpose or service hours change.