Results worth tracking

Choose the decision before the dashboard

Begin with the business question the campaign is meant to address. You might want suitable customers to request an estimate, existing customers to understand a new service or local visitors to check opening information. Choose one primary outcome and a small number of supporting signals. If the objective is vague, a dashboard can look impressive without telling you whether the work was useful.

Define what qualifies as an outcome before publishing. A qualified inquiry might mean a request from someone in your service area who needs an offer you actually provide. Keep that definition consistent through the campaign. Record completed bookings separately from inquiries, and revenue separately from bookings. These stages are related, but they are not interchangeable.

Keep metric definitions beside the numbers

Platforms do not always count behavior in the same way. A video view, impression or interaction may have a different definition across products and may change over time. Record the source and reporting period. Avoid combining metrics just because they share a similar label. Use the platform’s current documentation when a definition affects your decision.

Reach

An estimate of distinct accounts exposed to content within a platform’s reporting scope.

Impressions

The number of recorded displays; one account can contribute more than one.

Engagement rate

A selected set of interactions divided by a stated denominator, multiplied by 100.

Qualified inquiries

Contacts that meet your written relevance criteria; not every message qualifies.

Production time

Hours spent researching, creating, publishing, responding and reviewing.

If you calculate an engagement rate using reach, label it that way. A rate based on followers or impressions answers a different question. State which interactions you included. There is no single universal engagement rate that makes all channels and formats directly comparable.

Read the content of engagement

A reaction count can indicate attention, but it does not reveal why someone responded. Look at relevant questions, objections, requests for clarification and recurring language. Saves may suggest future usefulness; shares may suggest relevance to someone else. Neither proves buying intent. Treat the behavior as a clue to investigate rather than a substitute for a business outcome.

Review low-engagement content in context. A clear opening-hours update may help a small number of customers without attracting public replies. A broad entertainment post may receive attention from people outside your market. Assess the job the content was designed to do before deciding which one performed better.

Track pathways without claiming certainty

Consistent campaign links can help distinguish traffic sources. UTM parameters such as source, medium and campaign provide labels that compatible analytics tools can read. Use a simple naming convention and never put personal information in these parameters. Test the destination on a phone before publishing. A broken link can undermine both the customer journey and the report.

Google Analytics 4 can provide website event reporting when it is lawfully configured and appropriate for your circumstances. Google Search Console reports search-related information, not the full effect of a social campaign. Neither is installed by these educational pages. Choose tools based on your actual needs and meet applicable privacy and consent requirements before deploying them on your own site.

Attribution remains incomplete. People switch devices, reject tracking, share links privately and contact a business later through another route. A voluntary “How did you hear about us?” question adds context but relies on memory. Combine sources carefully and avoid counting the same inquiry twice. Describe observed associations rather than claiming the campaign caused every recorded sale.

Compare like with like where possible

Review similar reporting windows, formats and distribution conditions. Separate paid delivery from organic delivery. Note changes to the offer, audience, budget, availability and season. A campaign run during a holiday period may not be comparable to an ordinary week. Small samples can move dramatically because of one or two events.

If you test two approaches, change one important element where practical and write down the hypothesis first. Informal comparisons are useful for learning, but they are not automatically controlled experiments. Avoid declaring a winner from a small difference without enough evidence. Sometimes the most honest conclusion is that the result is inconclusive.

End with a decision you can act on

Keep the review short: objective, activity, observed outcome, time cost, limitations and next action. Include the questions people asked and any service problems the campaign exposed. If you calculate cost per qualified inquiry, state whether the cost includes advertising only or also production and owner time. Do not label a revenue-to-ad-spend ratio as profit.

Choose whether to continue, adapt, pause or investigate. A campaign with modest reach may still deserve another cycle if it answered important questions and produced relevant contacts at a manageable cost. A popular campaign may need revision if it attracted the wrong audience or created more support work than the business could handle.