Start with the question
Attribution is a way of assigning credit for an action to one or more marketing touchpoints. In a small business the practical question is often simpler: which activities appear to help suitable customers find and choose us? A useful answer combines the evidence available with an honest account of what that evidence misses.
A person might hear a recommendation, search for the business, read a guide and contact the team a week later. A report may show only the final recorded visit; the customer may remember the recommendation. Neither view automatically proves the full sequence. Ask what decision you need to make before choosing a reporting method.
Separate a source from an outcome
A source label describes a recorded route, such as a search visit or a campaign link. An outcome could be a useful enquiry, a completed booking or an agreed project. These are different events and may be captured in different systems. Check that a submission is genuinely delivered before using a button click as a substitute for an enquiry.
Define what you mean by a useful enquiry and who decides. For example, a service-area business may need the requested work and general location to fit its capacity. Use broad categories and avoid copying names, email addresses or the contents of messages into an analytics report.
Why reports can disagree
A search platform and a website analytics tool may count different events, cover different periods or use different definitions. Browser settings, consent choices, blocked scripts and processing delays can also affect what is recorded. A call or an in-person conversation may never appear in the website report. Compare like with like before assuming that a difference means one tool is broken.
Check dates, time zones, filters and whether the same domain and pages are included. Google notes that Search Console and Analytics measure different parts of the journey. A discrepancy is a reason to understand the setup, not an invitation to fill the gap with an invented figure.
One customer, several possible touchpoints
- RecommendationSomeone hears about the business
- SearchThey find a relevant page
- ConsiderThey read the information
- EnquireThey choose a contact route
Use a customer question carefully
Asking how someone heard about your business can add useful context if the question is genuinely relevant and optional where appropriate. Offer a small set of understandable choices and an option for not knowing or more than one route. Do not pressure someone to remember every touchpoint, and avoid storing more of their explanation than the business needs.
A customer may name a friend while the report labels a search visit. Both can be true: the recommendation created interest and the search helped them find the contact page. Record the distinction rather than overwriting one source to make it fit the other. Explain the use of personal information in your privacy information.
An illustrative decision
Imagine a team has limited time for two activities. A short guide is mentioned in several customer conversations, while a report also shows visits from a tagged newsletter. The team checks whether the article answers a recurring question and whether the newsletter sends readers to a working service page. It then decides whether to improve either asset or try a different message.
The example does not claim a real business achieved a result. It shows why qualitative feedback can sit alongside aggregated data. If only a handful of enquiries are involved, record the counts and the period rather than claiming a precise percentage proves one channel is responsible.
Make a proportionate review
Create a short table with activity, purpose, available signal, limitation and next action. For a search page, the signal might be impressions and relevant questions; for a newsletter, it might be a tested tagged link and replies; for a local event, it might be agreed follow-up conversations. Avoid collecting a new identifier merely to make a report look complete.
Review the table at a sensible interval and note what changed in staffing, offers or availability. Decide whether the evidence is sufficient to continue, adjust or pause an activity. Where measurement uses storage, advertising tools or personal data, check the current privacy rules and provider configuration for the actual implementation.
A three-source comparison exercise
For an imaginary enquiry, the team has three clues: the customer mentions a recommendation, the website report records a search visit and a newsletter link has its own campaign label. Write each as a separate observation with its source and limitation. The recommendation is self-reported; the search record describes a captured visit; the campaign label describes a link, not a confirmed sale.
Ask whether the observations relate to the same event before combining them. With a small group, linking records to an individual may be unnecessary or inappropriate. An aggregate comparison may be enough to decide whether a useful article should be updated or whether a destination page needs a clearer explanation. Check access, privacy information and the tools used for any data processing.
Record a decision such as 'keep the guide current and test the newsletter destination next month', along with what evidence would change that decision. If a source appears weaker after a tool change, first check the tracking and the time period. This prevents a technical reporting gap being mistaken for a sudden change in customer demand.
If a report labels a visitor's route as unknown, preserve that category rather than guessing. Changes to browser settings, consent, a link or a provider's reporting can affect the visible source. Record the uncertainty beside the count and compare the same definition over time; a smaller honest dataset is a better basis for discussion than a reconstructed customer journey.
Questions for your next report
What exactly is being counted? Which period and audience does the report cover? Do the numbers reflect completed actions or merely clicks? What journeys are missing, and have you checked the destination and contact route? Finally, what decision could change as a result of the evidence?
If you cannot answer those questions, improve the definition or simplify the report. A smaller, trustworthy set of signals is generally more useful than a detailed dashboard whose meaning is unclear. Share only the information needed by the people responsible for the decision and review access as roles change.
