How to measure the impact of social networks on brand recognition
Why ER and reach do not measure brand recognition
Reach and engagement show how many people saw the content and how they reacted to it. Recognition is the ability to recall a brand without visual contact with the post, through weeks or months. There is no direct proportion between these things. A post may gather hundreds of thousands of views due to appearing in recommendations, but if the recommendation audience does not match the target audience, brand recognition in that segment will not budge even a single percent.
A classic mistake is to take ER as is and rejoice in high figures. If an account has 1000 subscribers, and one post appeared in recommendations and gathered reach 100 000, the typical ER might show 1000% — a figure that explains nothing. Here a fairer formula applies:
ERR (Engagement Rate by Reach) = (engagement / reach) × 100%
For the same example, ERR will give objective 10% — real engagement relative to those who saw the post, and not relative to subscribers. The second common mistake is not cleaning reach from bots and dead accounts: 2–3% subscribers in most commercial profiles are inactive or fake records, and they distort all derived metrics.
Three levels of recognition and what affects them
Before linking social media metrics to recognition, it's necessary to understand that recognition itself is not a single number, but a hierarchy of three levels.
- Unaided recall — a person mentions the brand on their own, without a prompt, answering the question 'what companies in this category do you know'. The most difficult level to grow.
- Aided recognition — a person recognizes a name or a logo from the provided list. It grows faster, social networks influence it directly through frequency of exposure.
- Top of Mind — a brand that first comes to mind in a category. This is a goal, not an intermediate result: social networks influence it only through long-term, regular, and stylistically recognizable communication.
Social media metrics relate to these levels differently. Reach and frequency of exposure work on recognizability with a prompt — this is pure mechanics of repeated contact. Mentions, saves, and brand search queries are more closely linked to spontaneous recognizability, because they reflect active human behavior, not passive viewing. Social networks move Top of Mind the slowest and almost never — on their own.
Brand Lift: how to check the actual increase
Brand Lift is a standard study used in media advertising to check whether a campaign actually influenced brand perception, not just was displayed. The audience is divided into two groups: a test group that saw the ad or content, and a control group that did not. Both groups are asked the same questions about recognizability, and the difference in answers is the increase caused specifically by the campaign.
For example, if 78% of the test group recalls the brand, and only 50% in the control group does, the brand recall increase is 56% (relative change, not a percentage point difference). The same approach is applied to social networks in Brand Lift: the audience that was exposed to the brand's social media activity is compared with the audience that was not. This is the only method among those listed that distinguishes cause from coincidence — the other metrics only correlate with recall, but do not prove causal relationship.
Attribution problem in a multi-channel environment
Even if brand awareness is growing, it is difficult to prove that social networks are the cause, rather than search, contextual advertising, or PR publications. The 'last click' model, which still defaults in most analytics systems, works against the brand in this case: it attributes conversion to the last channel before the action, whereas social networks usually work on the first touch — a person sees a post, forgets about it, and then finds the brand through search after two weeks and 'buys' via contextual advertising.
| Attribution model | What it shows | Suitable for evaluating social networks |
|---|---|---|
| Last click | Channel directly before conversion | No — devalues the top of the funnel |
| First click | First touch channel | Partially — but ignores other touches |
| Positional (U-shaped) | Weights first and last touch more, others less | Yes — accounts for the role of social networks at the entrance to the funnel |
| Data-driven model (GA4) | Weights each channel according to an algorithm trained on real conversion paths | Yes, most accurate, but requires volume of data |
Practical takeaway — use positional attribution or the data-driven GA4 model together with growth in branded search traffic. The number of queries for the company name in Google or Yandex is an indirect but reliable indicator: it does not depend on who 'received' the conversion, and often correlates with social media activity with a lag of one to two weeks.
How to build a monitoring system: metrics and KPIs by business types
A scattered set of metrics does not provide a complete picture — a system is needed where social media indicators and recognition metrics regularly observe each other, not just occasionally. The working sequence looks like this:
- Establish a baseline recognition level through surveys (spontaneous, with hints, Top of Mind) before starting active work in social networks.
- Set up collection of ERR, saves, and mentions as key indicators of 'quality' contact, not just reach.
- Enable monitoring of brand search queries and Share of Voice relative to competitors.
- Repeat the recognition survey once a quarter or half-year and compare dynamics with the growth of indirect metrics.
- Conduct Brand Lift testing with a test and control group when launching major campaigns.
Share of Voice (SSoV) is calculated as the brand's share of mentions from all mentions in the category — (your mentions / mentions by all competitors) × 100%. This is not an absolute measure of brand recognition, but a metric for comparing with competitors: an increase in SSoV against a declining recognition rate in the entire category may mean that the brand is simply losing positions more slowly than others, rather than growing on its own.
The set of KPIs varies by stage and type of business:
- Launching a new brand on social media — reach, impressions, subscriber growth as a base, plus ER, retention, and mentions are mandatory: without the last three, reach only shows content delivery, not engagement.
- Local business (services, retail) — location-based mentions, brand searches in a specific region, share of voice in the local market.
- B2B and complex products — retention of expert content, mentions in professional communities, growth in brand traffic to the site from organic search.
- Mass consumer brand — regular Brand Lift, ERR by campaigns, comparison of spontaneous recognition before and after active periods.
Simple conclusion: numbers from the social media dashboard by themselves prove nothing — they only become meaningful when compared with direct measurements of recognition through surveys and Brand Lift. Companies that report only reach and ER are measuring activity, not results.
Why ER and reach do not measure brand recognition
Reach and engagement show how many people saw the content and how they reacted to it. Recognition is the ability to recall a brand without visual contact with the post, through weeks or months. There is no direct proportion between these things. A post may gather hundreds of thousands of views due to appearing in recommendations, but if the recommendation audience does not match the target audience, brand recognition in that segment will not budge even a single percent.
A classic mistake is to take ER as is and rejoice in high figures. If an account has 1000 subscribers, and one post appeared in recommendations and gathered reach 100 000, the typical ER might show 1000% — a figure that explains nothing. Here a fairer formula applies:
ERR (Engagement Rate by Reach) = (engagement / reach) × 100%
For the same example, ERR will give objective 10% — real engagement relative to those who saw the post, and not relative to subscribers. The second common mistake is not cleaning reach from bots and dead accounts: 2–3% subscribers in most commercial profiles are inactive or fake records, and they distort all derived metrics.
Three levels of recognition and what affects them
Before linking social media metrics to recognition, it's necessary to understand that recognition itself is not a single number, but a hierarchy of three levels.
- Unaided recall — a person mentions the brand on their own, without a prompt, answering the question 'what companies in this category do you know'. The most difficult level to grow.
- Aided recognition — a person recognizes a name or a logo from the provided list. It grows faster, social networks influence it directly through frequency of exposure.
- Top of Mind — a brand that first comes to mind in a category. This is a goal, not an intermediate result: social networks influence it only through long-term, regular, and stylistically recognizable communication.
Social media metrics relate to these levels differently. Reach and frequency of exposure work on recognizability with a prompt — this is pure mechanics of repeated contact. Mentions, saves, and brand search queries are more closely linked to spontaneous recognizability, because they reflect active human behavior, not passive viewing. Social networks move Top of Mind the slowest and almost never — on their own.
Brand Lift: how to check the actual increase
Brand Lift is a standard study used in media advertising to check whether a campaign actually influenced brand perception, not just was displayed. The audience is divided into two groups: a test group that saw the ad or content, and a control group that did not. Both groups are asked the same questions about recognizability, and the difference in answers is the increase caused specifically by the campaign.
For example, if 78% of the test group recalls the brand, and only 50% in the control group does, the brand recall increase is 56% (relative change, not a percentage point difference). The same approach is applied to social networks in Brand Lift: the audience that was exposed to the brand's social media activity is compared with the audience that was not. This is the only method among those listed that distinguishes cause from coincidence — the other metrics only correlate with recall, but do not prove causal relationship.
Attribution problem in a multi-channel environment
Even if brand awareness is growing, it is difficult to prove that social networks are the cause, rather than search, contextual advertising, or PR publications. The 'last click' model, which still defaults in most analytics systems, works against the brand in this case: it attributes conversion to the last channel before the action, whereas social networks usually work on the first touch — a person sees a post, forgets about it, and then finds the brand through search after two weeks and 'buys' via contextual advertising.
| Attribution model | What it shows | Suitable for evaluating social networks |
|---|---|---|
| Last click | Channel directly before conversion | No — devalues the top of the funnel |
| First click | First touch channel | Partially — but ignores other touches |
| Positional (U-shaped) | Weights first and last touch more, others less | Yes — accounts for the role of social networks at the entrance to the funnel |
| Data-driven model (GA4) | Weights each channel according to an algorithm trained on real conversion paths | Yes, most accurate, but requires volume of data |
Practical takeaway — use positional attribution or the data-driven GA4 model together with growth in branded search traffic. The number of queries for the company name in Google or Yandex is an indirect but reliable indicator: it does not depend on who 'received' the conversion, and often correlates with social media activity with a lag of one to two weeks.
How to build a monitoring system: metrics and KPIs by business types
A scattered set of metrics does not provide a complete picture — a system is needed where social media indicators and recognition metrics regularly observe each other, not just occasionally. The working sequence looks like this:
- Establish a baseline recognition level through surveys (spontaneous, with hints, Top of Mind) before starting active work in social networks.
- Set up collection of ERR, saves, and mentions as key indicators of 'quality' contact, not just reach.
- Enable monitoring of brand search queries and Share of Voice relative to competitors.
- Repeat the recognition survey once a quarter or half-year and compare dynamics with the growth of indirect metrics.
- Conduct Brand Lift testing with a test and control group when launching major campaigns.
Share of Voice (SSoV) is calculated as the brand's share of mentions from all mentions in the category — (your mentions / mentions by all competitors) × 100%. This is not an absolute measure of brand recognition, but a metric for comparing with competitors: an increase in SSoV against a declining recognition rate in the entire category may mean that the brand is simply losing positions more slowly than others, rather than growing on its own.
The set of KPIs varies by stage and type of business:
- Launching a new brand on social media — reach, impressions, subscriber growth as a base, plus ER, retention, and mentions are mandatory: without the last three, reach only shows content delivery, not engagement.
- Local business (services, retail) — location-based mentions, brand searches in a specific region, share of voice in the local market.
- B2B and complex products — retention of expert content, mentions in professional communities, growth in brand traffic to the site from organic search.
- Mass consumer brand — regular Brand Lift, ERR by campaigns, comparison of spontaneous recognition before and after active periods.
Simple conclusion: numbers from the social media dashboard by themselves prove nothing — they only become meaningful when compared with direct measurements of recognition through surveys and Brand Lift. Companies that report only reach and ER are measuring activity, not results.
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
