KPI for brand social media: which metrics are truly important
Reach is growing, but there are no applications — where to look for the problem
This is the first situation that almost everyone encounters who has been analyzing brand social media analytics for more than a month. Reach and impressions numbers are going up, content is gathering reactions, but calls and applications are still non-existent. The solution lies in the fact that metrics describe different floors of the funnel, and growth on the upper floor does not guarantee movement on the lower ones.
All indicators can be conditionally divided into touch metrics and choice metrics. Touch includes reach, impressions, views, brand mentions: the audience has seen the content but has not decided anything. Choice includes clicks on links, transitions in Direct, saves, additions to favorites, applications: the audience has taken an action that indicates intent. The problem with most reports is that they combine these two levels into one table and rejoice in overall growth, although the growth may be entirely due to touch, while choice may be declining.
If there is reach but no applications, you need to look at the intermediate layer: CTR on the link in the profile header, the number of transitions to the site, the share of exits on these transitions. 1000 transitions with an exit rate of 90% is worse than 200 transitions where people browse through the cards and add the product to the cart. A beautiful traffic figure without quality of this traffic — is a classic self-deception.
How to choose one main KPI to not drown in numbers
A dashboard with thirty metrics does not help make decisions — it hides them. A working system is built in reverse order: first, a business goal is formulated, then exactly one main indicator is selected for it and 3–5 supporting metrics, which explain the dynamics of the main one. Everything else is background, which is reviewed once a quarter, not weekly.
| Goal | Main KPI | Supporting metrics |
|---|---|---|
| Brand Launch | Reach of unique audience | ER, brand mentions, brand searches |
| Community | Repeat interactions | ER, comments, share of regular audience |
| Traffic | Clicks on the site | CTR, bounce rate, depth of view |
| Leads | Number of requests (CR) | CPL, lead quality, share of target calls |
| Recognizability | Brand searches / mentions | Reach, survey results on recognition |
Main KPIs are not observed at a given moment, but rather in trends over weeks. One good week doesn't prove anything: a contest, a viral video, or a seasonal surge can easily give a one-time jump. Only sustained results become visible over a period of 6–8 weeks, when random fluctuations are averaged out.
Microconversions and macroconversions: why you can't measure only sales
Many brands only track macroconversions — order, request, payment — and are surprised why the numbers aren't moving, even though the content clearly improved. The problem is that there are several intermediate steps between first contact and purchase, and if these aren't tracked, it's impossible to understand where the audience is actually being lost.
- Microconversions — subscription to newsletter, downloading a checklist, saving a post, going to direct messages. These are signals of intent that precede a purchase.
- Macroconversions — application, order, payment. Direct business result.
- Behavioral metrics — adding to favorites, viewing product card, repeat visit to profile. Show how much the audience has 'warmed up'.
Ignoring micro-conversions is a mistake, because it is precisely these that allow you to see the problem before it reflects in revenue. If the number of check-list downloads has decreased despite the same reach, this is an early warning: the content is losing relevance for the audience that is already ready for the next step. Waiting until this affects sales — is to lose at least a month to fix it.
How to honestly calculate ROMI and not fool yourself
ROMI is calculated by the formula: ((revenue – cost) × 100) ÷ cost. A result above 100% means that the investment not only paid off, but also brought additional income beyond the costs. The formula is simple, but self-deception arises not in the mathematics, but in what is recorded in 'revenue'.
A common manipulation is to count all sales to customers who have ever seen ads in social networks, including those who would have come anyway without them. A fair calculation requires attribution: revenue is counted only for conversions that went through a tracked path — a promo code, UTM tag, a separate phone number or form. The second point is that expenses should include not only the advertising budget, but also the cost of content production, the team's work, and analytics tools. ROMI calculated only based on media budget without accounting for production always looks better than it actually is.
A useful benchmark for checking the realism of figures is the typical conversion rate by niche. In e-commerce with warm traffic, landing page CR stays within a few percent; in B2B with a high average order value, a normal rate might be 0,5–1%. If a report shows a conversion many times higher than these ranges without an obvious explanation, it is likely that there is duplication or incorrect attribution somewhere in the calculation.
What speaks to the quality of the audience, not the amount of activity
The number of subscribers is the most deceptive metric in SMM, because it is easy to inflate it with an event unrelated to the product. A rapid audience growth on 30% due to a contest or marathon, coupled with a simultaneous increase in unsubscribes after its end, almost always means that non-target users came, not the target audience, but gift hunters. Formally, 'activity' was there, but in reality, the business gained nothing.
Audience quality metrics are those that show the depth and repetitiveness of interaction, not a one-time spike: the share of comments with substantive text, not emojis; the percentage of subscribers who interact more than once per month; the ratio of new and repeat inquiries from social networks. ER calculated as (likes + comments + shares + saves) / reach or subscribers × 100% also gives different pictures depending on what is taken in the denominator — one should fix one methodology and not change it between reports, otherwise the trend ceases to be comparable.
Common mistakes when building a KPI system
- Mixing touchpoint metrics and selection metrics in one report without separation — this masks the decline in conversion rate with the growth in reach.
- Choosing the main KPI for the reporting tool, not for the business goal — if analytics can only count likes, this is not a reason to make likes a KPI.
- Ignoring micro-conversions and waiting for a signal only from sales — time is lost on early problem diagnosis.
- Calculating ROMI without accounting for production costs and without correct revenue attribution.
- Evaluating subscriber growth in isolation from the source of that growth — not distinguishing organic growth from competitive.
A KPI system that actually works does not try to cover everything at once. It is built from top to bottom: goal — one main indicator — several supporting metrics that explain its dynamics — regular monitoring of trends, not one-time figures. Everything that does not help answer the question 'are we moving towards the goal' can be reviewed once a month for a general picture, but should not be included in the weekly report.
Reach is growing, but there are no applications — where to look for the problem
This is the first situation that almost everyone encounters who has been analyzing brand social media analytics for more than a month. Reach and impressions numbers are going up, content is gathering reactions, but calls and applications are still non-existent. The solution lies in the fact that metrics describe different floors of the funnel, and growth on the upper floor does not guarantee movement on the lower ones.
All indicators can be conditionally divided into touch metrics and choice metrics. Touch includes reach, impressions, views, brand mentions: the audience has seen the content but has not decided anything. Choice includes clicks on links, transitions in Direct, saves, additions to favorites, applications: the audience has taken an action that indicates intent. The problem with most reports is that they combine these two levels into one table and rejoice in overall growth, although the growth may be entirely due to touch, while choice may be declining.
If there is reach but no applications, you need to look at the intermediate layer: CTR on the link in the profile header, the number of transitions to the site, the share of exits on these transitions. 1000 transitions with an exit rate of 90% is worse than 200 transitions where people browse through the cards and add the product to the cart. A beautiful traffic figure without quality of this traffic — is a classic self-deception.
How to choose one main KPI to not drown in numbers
A dashboard with thirty metrics does not help make decisions — it hides them. A working system is built in reverse order: first, a business goal is formulated, then exactly one main indicator is selected for it and 3–5 supporting metrics, which explain the dynamics of the main one. Everything else is background, which is reviewed once a quarter, not weekly.
| Goal | Main KPI | Supporting metrics |
|---|---|---|
| Brand Launch | Reach of unique audience | ER, brand mentions, brand searches |
| Community | Repeat interactions | ER, comments, share of regular audience |
| Traffic | Clicks on the site | CTR, bounce rate, depth of view |
| Leads | Number of requests (CR) | CPL, lead quality, share of target calls |
| Recognizability | Brand searches / mentions | Reach, survey results on recognition |
Main KPIs are not observed at a given moment, but rather in trends over weeks. One good week doesn't prove anything: a contest, a viral video, or a seasonal surge can easily give a one-time jump. Only sustained results become visible over a period of 6–8 weeks, when random fluctuations are averaged out.
Microconversions and macroconversions: why you can't measure only sales
Many brands only track macroconversions — order, request, payment — and are surprised why the numbers aren't moving, even though the content clearly improved. The problem is that there are several intermediate steps between first contact and purchase, and if these aren't tracked, it's impossible to understand where the audience is actually being lost.
- Microconversions — subscription to newsletter, downloading a checklist, saving a post, going to direct messages. These are signals of intent that precede a purchase.
- Macroconversions — application, order, payment. Direct business result.
- Behavioral metrics — adding to favorites, viewing product card, repeat visit to profile. Show how much the audience has 'warmed up'.
Ignoring micro-conversions is a mistake, because it is precisely these that allow you to see the problem before it reflects in revenue. If the number of check-list downloads has decreased despite the same reach, this is an early warning: the content is losing relevance for the audience that is already ready for the next step. Waiting until this affects sales — is to lose at least a month to fix it.
How to honestly calculate ROMI and not fool yourself
ROMI is calculated by the formula: ((revenue – cost) × 100) ÷ cost. A result above 100% means that the investment not only paid off, but also brought additional income beyond the costs. The formula is simple, but self-deception arises not in the mathematics, but in what is recorded in 'revenue'.
A common manipulation is to count all sales to customers who have ever seen ads in social networks, including those who would have come anyway without them. A fair calculation requires attribution: revenue is counted only for conversions that went through a tracked path — a promo code, UTM tag, a separate phone number or form. The second point is that expenses should include not only the advertising budget, but also the cost of content production, the team's work, and analytics tools. ROMI calculated only based on media budget without accounting for production always looks better than it actually is.
A useful benchmark for checking the realism of figures is the typical conversion rate by niche. In e-commerce with warm traffic, landing page CR stays within a few percent; in B2B with a high average order value, a normal rate might be 0,5–1%. If a report shows a conversion many times higher than these ranges without an obvious explanation, it is likely that there is duplication or incorrect attribution somewhere in the calculation.
What speaks to the quality of the audience, not the amount of activity
The number of subscribers is the most deceptive metric in SMM, because it is easy to inflate it with an event unrelated to the product. A rapid audience growth on 30% due to a contest or marathon, coupled with a simultaneous increase in unsubscribes after its end, almost always means that non-target users came, not the target audience, but gift hunters. Formally, 'activity' was there, but in reality, the business gained nothing.
Audience quality metrics are those that show the depth and repetitiveness of interaction, not a one-time spike: the share of comments with substantive text, not emojis; the percentage of subscribers who interact more than once per month; the ratio of new and repeat inquiries from social networks. ER calculated as (likes + comments + shares + saves) / reach or subscribers × 100% also gives different pictures depending on what is taken in the denominator — one should fix one methodology and not change it between reports, otherwise the trend ceases to be comparable.
Common mistakes when building a KPI system
- Mixing touchpoint metrics and selection metrics in one report without separation — this masks the decline in conversion rate with the growth in reach.
- Choosing the main KPI for the reporting tool, not for the business goal — if analytics can only count likes, this is not a reason to make likes a KPI.
- Ignoring micro-conversions and waiting for a signal only from sales — time is lost on early problem diagnosis.
- Calculating ROMI without accounting for production costs and without correct revenue attribution.
- Evaluating subscriber growth in isolation from the source of that growth — not distinguishing organic growth from competitive.
A KPI system that actually works does not try to cover everything at once. It is built from top to bottom: goal — one main indicator — several supporting metrics that explain its dynamics — regular monitoring of trends, not one-time figures. Everything that does not help answer the question 'are we moving towards the goal' can be reviewed once a month for a general picture, but should not be included in the weekly report.
Recent articles
MegaFon, BMW, VTB and more 240+ companies. See how it looks in the showcase reel:
Our branding agency is approached to order:
Branding agency services
