What are the 5 main KPIs for social media?

Reach, engagement rate, saves and shares, follower growth rate, and conversions attributed to social. Most agencies add a sixth, publishing consistency, because it is the only one fully inside their control. In 2026, 44% of agency clients named conversions their single most important outcome metric.

What are some good KPI examples for a client retainer?

Good ones name a number and a date: non-follower reach above 40% of total reach by 31 March, engagement rate held above the 0.30% platform median, four posts published every week. Bad ones name a metric and nothing else, like 'grow engagement'.

How many KPIs should a social media report include?

Six or fewer on the client-facing page. Every KPI you show is a promise you have to defend, so twenty metrics with a 30% chance each of moving the wrong way produces six red numbers a month against the same underlying work that six KPIs would show as roughly two.

How do you set a social media KPI target?

Take the benchmark for the client's follower tier, take the account's own trailing 90 day median, then add the delta your specific intervention can plausibly buy. Never a round number. A target of 0.5% is a guess; 0.47% derived from a 0.41% baseline is an argument you can defend.

Can you change a social media KPI mid-engagement?

Yes, if you change it on a date that was in the contract before the number went red. Put a named quarterly review in the scope, bring the benchmark movement as evidence rather than the shortfall, and show the delta you did deliver. Renegotiating after a miss reads as goalpost moving.

Social Media KPIs: The Six That Survive Client Scrutiny

In 2026, agencies that set realistic KPIs during client onboarding retained clients 15 to 20 percentage points better than the industry average (Focus Digital, 2026). Not smarter KPIs. Realistic ones, agreed early.

Every article ranking for this query is a list of twenty metrics with a definition each. That list is not the thing a client challenges you on. They challenge the number sitting next to the metric, where you got it, and what happens when you miss it. A KPI without a target is a definition. A target without a benchmark is a guess.

Key Takeaways

Realistic KPIs agreed at onboarding are worth 15 to 20 percentage points of retention (Focus Digital, 2026).

Six client-facing KPIs, not twenty. Every metric you report is a promise you have to defend monthly.

A target is a benchmark plus a committed delta, never a round number. Round numbers are the tell that you invented it.

Five of the six stay fixed for the life of the retainer. Only the conversion KPI moves, and only when the client's business goal moves.

Why is a twenty KPI dashboard a retention risk?

Because it multiplies your exposure without improving the work. In 2026, 84% of agency clients said clear visual representation of data mattered most to them in a report, ahead of campaign access at 48% and actionable metrics at 42% (AgencyAnalytics, 2026). Clarity beat volume, and it was not close.

Run the arithmetic once and you never build a twenty metric dashboard again. Assume any given metric has a 70% chance of moving in the direction you want in any given month, which is generous. With twenty KPIs on the page, you can expect six red numbers every single month. With six KPIs, you can expect two.

Same account. Same work. Three times the objections to answer on the call.

There is a second cost, and it is worse. In 2026, 55% of agency marketers said they were overwhelmed by the data they collect, and 28% admitted that when they report results they do not analyse the why (Funnel, 2026). Every metric you add is a metric you will not have time to explain. So the real question is not which metrics matter. It is narrower: which six are you prepared to be judged on for twelve months?

The six KPIs that survive a client challenge

These six survive because each one is evidence of something a client already believes is worth money. In 2026, Quid's benchmark report put the median Instagram engagement rate at 0.30% per post against followers, down from 0.36% the year before (Quid, 2026). A falling denominator is exactly why each KPI below needs its source stated.

KPI	What it actually proves	2026 benchmark	Where the number comes from
Reach rate (reach over followers)	The work is still being distributed	5.50% at 10K to 50K, 3.50% at 100K to 1M	Socialinsider, 872,075 posts to Aug 2026
Engagement rate	Content quality, adjusted for size	0.45% per follower, 0.30% median, 3.54% per reach	Three publishers, three denominators
Saves plus shares per post	The content was worth keeping or passing on	Shares rose to 45 per post, the only metric that grew	Socialinsider, 70M posts
Follower growth rate	Trajectory, normalised for account size	17.20% annual at 10K to 50K, 11.25% at 100K to 1M	Socialinsider, full year 2025 data
Publishing consistency	The agency did the work it was paid for	1.30, 3.69 or 5 posts per week	Metricool, Quid and Socialinsider, all disagreeing
Conversions attributed to social	The outcome the client is actually buying	Organic social sends 1.47% of all site traffic	Semrush, 50,000 websites

Source: Socialinsider, Quid and Semrush, 2026.

Look hard at the engagement row. Socialinsider put Instagram at 0.45% for the first half of 2026, Quid put the median at 0.30%, and Metricool reported 3.54%, because Metricool divides by reach and the other two divide by followers. Quote a benchmark without naming the denominator and a sharp client will find one of the others. The calculation is in the Instagram engagement rate guide, and the tier tables are in what is a good Instagram engagement rate.

The saves row hides the best argument in the table. Between 2024 and 2025 Instagram's average likes per post fell 15% and comments fell 20%, while views rose 29% and shares rose 12%. The metrics agencies have historically reported are the ones dying. Saves and shares covers why, and the reach row depends on the split explained in reach versus impressions.

This walkthrough is a decent sanity check on which metrics earn their place.

How do you set the target number?

Benchmark plus committed delta. In 2026, Metricool studied 24,364,803 Instagram posts and found brands published 24.04% more content year over year while total interactions rose only 19.25% (Metricool, 2026). Per post, engagement fell. Any target that assumes flat conditions is already wrong.

Here is the derivation, run properly. Client sits at 24,000 followers, so the 10K to 50K tier. Their trailing 90 day median engagement rate is 0.41%. The tier is drifting down roughly in line with the platform. Your intervention is a format shift from single images to carousels, and Metricool measured carousels generating nine times more saves than single images.

So the target is not 0.5%. It is 0.47% by 31 March, built from a 0.41% baseline, a benchmark that is falling about 17% a year, and a format change with a measured effect size. When the client asks where 0.47% came from, that sentence is the answer.

Notice what just happened. Holding 0.41% flat through a declining year is itself a win, and you can say so in advance instead of arguing it after the fact.

Round numbers are the tell. 5% growth, 1,000 new followers, 10% engagement lift: nobody arrives at those from data. Clients who have hired agencies before know it, and the ones who have not will learn it from the first quarter you miss.

One more discipline. Name the benchmark publisher in the scope of work and never switch it. Median Instagram posting frequency in 2026 is 1.30 posts a week per Metricool, 3.69 per Quid, and around five per Socialinsider. All three are honest, and their samples differ. Swapping to whichever one flatters you this quarter is the fastest way to lose the room.

Which KPIs belong in the client report and which stay internal?

The six above go to the client. Everything else stays in your workspace. With 69% of agencies reporting monthly in 2026 and 84% of clients prioritising visual clarity (AgencyAnalytics, 2026), the monthly page has room for outcomes and nothing else.

The distinction is simple. Client KPIs are outcomes. Internal metrics are levers. Show a client a lever and they will start pulling it, which is how you end up defending your posting schedule instead of your results.

Keep these in the internal view: format mix performance, posting time tests, hashtag splits (Metricool found hashtagged posts took 31.70% fewer views), caption length tests, comment response time, draft rejection rate, and hours logged per client. Every one of them informs your decisions. None of them belongs on a page the client's CFO might read.

There is a commercial reason too. Hours per client is the number that tells you whether the retainer is priced right. Would you show a client your margin? Then do not show them the inputs to it. Report structure section by section is covered in building a client social media report, so we will not repeat it here.

The one KPI that changes when the business goal changes

Five of the six are fixed. The sixth is bound to whatever the client is currently trying to do. In 2026, clients ranked conversions first at 44%, then leads at 22%, ROI at 12% and revenue at 11% (AgencyAnalytics, 2026). Those are four different sixth KPIs, not four names for one.

A client in launch mode wants qualified leads. The same client eighteen months later, with a full pipeline and a retention problem, wants repeat purchase influenced by social. If your sixth KPI is still leads, you are reporting against a goal the client abandoned two quarters ago and has not told you about.

Sprout Social found 68% of marketing leaders define social ROI by engagement, while 65% say what they actually want is a direct connection between social campaigns and business goals (Sprout Social, 2025). That gap is the whole job. Attribution reality is covered in proving social media ROI.

Size the sixth KPI honestly, though. Semrush's study of more than 50,000 websites found organic social sent just 1.47% of total website traffic in 2025, down 8.86% year over year (Semrush, 2026). If your conversion target quietly assumes social is a primary traffic channel, you have promised something the wider web is not delivering.

Ask the question at every quarterly review, in these words: what does the business need more of in the next ninety days? Then change one KPI, and only one.

How do you renegotiate a target without moving the goalposts?

You renegotiate on a date, not on a miss. In 2026 agencies named budget cuts their top churn driver at 42% (AgencyAnalytics, 2026), while clients surveyed inside the same twelve months named dissatisfaction with delivery and dissatisfaction with value, both at 61% (Setup, 2026). Agencies blame the economy. Clients blame the work. A target changed the week after it went red simply confirms the client's version.

Four things make the conversation survivable. Put a named review date in the scope of work before month one, so the revision is a scheduled event rather than a reaction. Bring the benchmark movement first, with the source, before you mention your own number. Show the delta you did deliver against the baseline, not against the target. And propose the new target with the same derivation you used the first time.

The order matters more than the content. Lead with your shortfall and the client hears an apology. Lead with a 17% decline in the platform median and your account holding flat inside it, and the client hears an analyst.

One structural fix helps here more than any script. Instagram's native insights cover roughly a recent 90 day window, which means the twelve month comparison a renegotiation depends on does not exist unless somebody stored it. An agency workspace that snapshots each client account daily keeps that history, and the alternative is a spreadsheet one person maintains and nobody else can reconstruct.

The renewal conversation is where all of this lands, and running the quarterly business review covers how to structure it.

Where to go next

Two posts pick up directly from here. Proving social media ROI covers what happens when the sixth KPI has to survive a finance review. Running the quarterly business review covers the meeting where targets get revised.

And if you want the benchmark numbers to hand before you set anything, what is a good Instagram engagement rate has the tier tables.