Social media KPIs: how to choose them by goal

A social media KPI is a key performance indicator that connects a goal to a measure and a decision. Choose it from the outcome you need, the base that defines it, and the source that records it.

Define the decision before choosing a metric

A metric counts something that happened: impressions, clicks, saves, or enquiries. A KPI is the metric you connect to an outcome and review when making a decision. If the team needs more bookings, for example, follower growth alone cannot show whether social media helped generate bookings.

Write the goal as an observable change: “receive workshop enquiries from a social campaign” is more useful than “improve social.” Then specify what counts as an enquiry, which accounts and content are included, the period, and the decision that depends on the result.

Choose one primary KPI and, when useful, one or two diagnostic measures. The primary KPI follows the outcome; secondary measures help explain its movement. A larger dashboard can support exploration, but the review should begin with the numbers that answer the agreed question. Use that criterion for the agreed campaign and period; revisit it if the audience, offer, or channel changes.

Match each source to the event. Native platform insights describe actions within a network. Website analytics records sessions and web events. A booking system or CRM confirms qualified enquiries and sales. A source can be right for one field and unsuitable for another.

Connect the goal to a metric and an action

Swipe the table to see all columns.

Reach and impressions answer different questions. Reach estimates how many accounts or people a platform reached according to its own definition; impressions count how often it showed the content. If the goal is to reach a local community, add the target area when the report supports it and separate organic from paid distribution.

For interest, choose the action the content is meant to prompt. If you want people to keep a guide, measure saves. If you want conversations, define which kind of comment counts as a relevant question. Reactions can add context, but by themselves they do not confirm that someone understood an offer or intends to buy.

To measure visits to a page, compare a platform’s outbound clicks with sessions that reach the site through tagged links. Google Analytics uses UTM parameters to identify source, medium, and campaign and show those values in traffic acquisition reports. Keep the naming consistent and save the campaign URL with the reporting period. Google Analytics guide to tagging campaigns. For network-specific tagging steps, see how to use social media UTM tracking.

For enquiries or sales, agree on a definition with the commercial team before the campaign begins. “Form submitted” may include spam or requests the business cannot serve. If quality matters, record a verifiable stage in the CRM, such as a qualified enquiry or confirmed booking, and keep the platform’s reported conversions separate.

A worked example with fictional data

Loma Taller is a fictional ceramics studio promoting an autumn course. It compares two organic posts with similar formats, one per period, on the same network and in the same report. Each post is measured over its first 30 days after publication.

Accounts reached2,400Second post

First post

1,900

Second post

2,400

Impressions5,100Second post

First post

3,600

Second post

5,100

The second post reached more accounts and accumulated more exposures. To decide whether the campaign built local awareness, the team checks how much of that reach came from the studio’s target area.

The studio also wants people to save a post listing the materials they need. The first post had 41 saves; the second had 34. Dividing each post’s saves by that same post’s reach gives rates of 2.16% and 1.42%. The save count fell, and so did the rate. Looking only at reach would hide that signal. Looking only at the rate would hide how many saves each version received. This formula is specific to the example, not a benchmark for other brands.

Finally, the second post’s link received 120 outbound clicks in the social network, while web analytics recorded 86 sessions with the campaign tag and the CRM listed four qualified enquiries during the same period without joining them to those sessions. These are three counts from different sources and with different units. The team checks that the tag survived redirects and that it applied the enquiry definition consistently before changing the content. It does not automatically divide four by 120: those totals do not prove that they describe the same people or establish a conversion rate.

These figures are instructional and fictional. They do not describe client results or set a target or industry reference.

Keep rates, counts, and definitions together

A rate helps compare periods or posts with different exposure; an absolute count shows action volume. Keep both, along with the base and formula. For example: “34 saves; 2,400 accounts reached; 34 ÷ 2,400 × 100 = 1.42%.” Saying only “1.42% engagement” is ambiguous unless you state which actions and denominator it includes.

Metric definitions also differ across reports. LinkedIn documents impressions, members reached, reactions, comments, and other actions separately; its organization engagement statistic combines organic clicks, likes, comments, and shares over impressions. That describes LinkedIn’s metric, not a formula shared by every network. Member post statistics and organization share statistics.

Check these differences before comparing:

• Unique reach: keep each network and reporting level separate. Adding reach by post or platform does not automatically count unique people across a campaign; audiences may overlap.
• Time windows: record start date, end date, time zone, and attribution window. LinkedIn statistics can return a post’s lifetime data when no period is specified. Compare equivalent periods and content scopes.
• Attribution: a UTM records a tagged source for a visit in website analytics. It does not prove that a post caused a purchase. A social network, a website, and a CRM may credit outcomes under different rules and windows.
• Missing data and distribution: distinguish zero from a value that is unavailable. Note changes in spend, collaborations, promotions, and tracking failures alongside the result.

TikTok Studio, for example, separates account and post analytics; TikTok also says some features can vary by region and app experience. Check which fields the account’s report provides before promising a specific metric. TikTok’s official creator tools.

Make the measurement agreement ready for review

Use the English social media KPI CSV template to record one goal per row: intended outcome, exact definition, base, period, source, decision, and limitations. Edit it in a spreadsheet. Use aggregated data; do not add names, email addresses, phone numbers, or other personal information.

Complete this brief before the reporting period begins:

Prepare a measurement agreement

Edit the bracketed fields for your brand.Business goal: [the outcome that matters] Networks, accounts, and content included: [scope] Primary KPI: [exact name and definition] Numerator, denominator, and unit: [include the absolute count and rate if relevant] Period, time zone, and attribution window: [dates and rule] Source for each field: [native analytics, web analytics, CRM, or bookings] Diagnostic measures: [only those that help explain the KPI] Agreed decision: [what to review, keep, change, or test] Limitations: [missing data, audience overlap, spend, windows, attribution] Separate observations from hypotheses. Leave unavailable values blank and use aggregated information without names, email addresses, or phone numbers.

To present the period to a client, use the social media report template. If you need to define a rate, see how to calculate social media engagement. If you plan to compare variations, see how to test social media content. For related topics, browse HeyMark’s social media guides.

To calculate a campaign’s financial return, use the social media ROI formula and template. State the included costs and attribution rule before comparing percentages.

Sources and scope

The linked official Google Analytics, LinkedIn, and TikTok documents support claims about report fields and measurement behavior. The selection method, Loma Taller case, and measurement brief are editorial work. The case data are fictional and do not provide a benchmark. The cover is an AI-generated illustration.