All posts
Brand & Content

How to measure brand awareness: 3 numbers for small brands

Small brands can skip impressions. Three numbers covering memory, branded search and sales show whether customers remember you after a campaign.

By 126Insights EditorialNovember 16, 2026 · 7 min read
Close-up of two hands holding a smartphone, one finger tapping the on-screen keyboard against a dark blurred background
Photo by freestocks on Unsplash

In mid-November, a lot of Vietnamese marketing teams are staring at the same spreadsheet: Tet budget, posting calendar, campaign concept. (Tet is the Lunar New Year, the biggest retail and gifting season of the year in Vietnam.) Few of them write down the question that will come back in March: after Tet, do customers remember us better? To answer it, you need a baseline measured before Tet, and that means collecting it in the next few weeks.

Search "how to measure brand awareness" and you get long lists. Zapier lists 9 key metrics, TheCMO offers 10 strategies and metrics, and LocalIQ gives 11 metrics and methods. Some Vietnamese guides file website traffic, social engagement and ad impressions under "brand awareness". 126Insights sees it differently. Those numbers tell you customers have seen you. They do not tell you whether customers remember you. A small brand needs three layers, one number per layer, measured the same way every period. And the cheapest number sits in the search box.

Do impressions measure brand awareness?

Vinabrand, a Vietnamese branding publication, puts it bluntly: "impressions, reach or views are not brand awareness." Those figures describe exposure. In Vinabrand's words, measuring brand awareness means measuring the result of memory, not just the level of media contact.

Think of a wedding. You shake hands with thirty friends of the bride and groom. A week later you run into one of them at the supermarket. The face looks familiar, but you may not come up with the name. Impressions are the handshakes. Finding a face familiar is recognition: you know it when prompted. Producing the name without seeing a name or logo is recall: you remember unprompted. Vinabrand calls these "two different memory tasks". When a customer types into a search box, no logo is there to remind them, so only recall brings them to you.

So for your brand: how many people can name you when you are not in the room?

How to measure brand awareness for a small brand: three layers, three numbers

The three layers 126Insights proposes are memory (do customers recall your name unprompted), behavior (does anyone type your name to look for you) and business (is money easier to earn). Keep one main number per layer. The reason is simple: a small team has nobody to track 11 metrics, and one number measured steadily for four quarters is worth more than ten numbers measured differently each quarter.

Memory layer: do customers recall your name unprompted?

Run a small survey inside your actual target group. Do not send it to your own followers, because they already know you. Suppose you sell roasted coffee. The first question can be just: "When you think of roasted coffee brands, which ones come to mind?" No list, no logo hints.

The formula from Vinabrand: unaided awareness equals the number of people who name the brand on their own, divided by the total valid respondents, times 100%. Count the people who name you first and you have top-of-mind. Vinabrand gives an illustration with made-up numbers: 400 respondents, 60 name brand X first, so X's top-of-mind is 15%. Because one person can name several brands, unaided awareness across the brands in a category can add up to more than 100%.

What about a new brand? A piece on Brands Vietnam argues that new brands should prioritize aided awareness (recognition when prompted) and established brands should prioritize unaided. 126Insights half agrees. Ask both in the same questionnaire: the unprompted question first, the question with a list of names second, so the list does not skew the first answer. The main number stays the unaided rate. While it sits near zero, the recognition rate shows the team it is inching up, so nobody gives up.

On timing, Sao Kim writes that brand awareness "can be measured quarterly". If you repeat an existing signal this Tet, as 126Insights suggests in its piece on keeping one brand sound for Tet 2027 music videos (in Vietnamese), your March measurement will tell you whether that signal stuck in customers' memory.

Bàn tay cầm bút bi phía trên bìa kẹp hồ sơ màu xanh, nền phía sau nhòe

Photo by Phil Hearing on Unsplash

Behavior layer: is anyone searching for your name?

This is the cheapest number. According to Attest, branded search volume shows how many people "actively remember your brand and seek it out directly". Nobody types a name they do not remember. Attest suggests tracking it with Google Analytics and Search Console, both free. The main number for this layer is the monthly count of searches containing your brand name. Google Trends adds the direction: per Attest, it shows whether interest is rising, flat or falling.

Brandwatch adds another source, direct traffic in Google Analytics, which it says gives an overview of current awareness. Brandwatch also warns that a generic brand name muddies the count. Suppose your shop is called "Moc" (Vietnamese for "raw wood" or "rustic"): your searches would mix with furniture, cafes and everything else. That is a good reason to think about this early, when naming a brand so it does not collide with others (in Vietnamese).

What follows is a suggestion from 126Insights. If you sell mostly on marketplaces, Google search volume does not tell the whole story. On the same day each month, type your name into the Shopee and TikTok search boxes and screenshot the autocomplete suggestions. In our view, the day your name starts appearing after the first few letters is a sign customers are looking for you. A screenshot does not replace a count, but kept steadily it becomes a diary you can compare over time.

If you have a budget for social listening, you can add share of voice (SOV). By the formula Kompa.ai published on Brands Vietnam, SOV equals mentions of your brand divided by total mentions of all brands, times 100%. Another piece on Brands Vietnam notes this KPI "does not consider quality". In its basic form SOV only counts mentions; to know whether they are praise or complaints you need further analysis.

Measuring brand performance in money: the business layer

Vinabrand makes a point that is easy to miss: high awareness does not prove that customers like you, consider you or will buy. Hence the third layer. Sao Kim proposes measuring in three tiers: output (work done), outcome (such as unaided awareness and purchase consideration) and impact. Long-term impact includes brand position, pricing power and lower selling costs. Also per Sao Kim, business impact "needs to be tracked monthly or by season", and in its section on common mistakes it warns: "Measuring brand impact after one month is too early."

A small brand should pick one of two numbers, depending on how it sells. If you sell on marketplaces or your own website, track advertising cost per order, the closest way to see "lower selling costs". If you often have to discount, track the share of orders sold at list price, which is close to the pricing power Sao Kim describes: raising prices while customers keep buying. This is how 126Insights turns two Sao Kim indicators into numbers you can read from sales reports you already have. It also lets you answer the budget question in the right order, as the piece on what percentage of revenue marketing should cost (in Vietnamese) discusses.

Three traps that spoil awareness numbers

The first trap is treating NPS (Net Promoter Score, a customer loyalty survey) as an awareness metric. Sao Kim lists unaided awareness and NPS as two separate indicators. To know whether customers recall your name unprompted, the number to watch is still the unaided rate from the memory layer.

The second trap is a survey with a prize. This is 126Insights' own judgment: an attractive prize pulls in people who answer for the prize and are not in your target group. If the questionnaire carries your logo, respondents have seen your name before being asked, and the unaided question loses its meaning. Send the survey under a neutral name.

The third trap is changing the question between two measurements. If this quarter you ask "which roasted coffee brands come to mind" and next quarter "which Vietnamese coffee brands do you know", the two numbers can no longer be compared. Keep the question, the respondent group and the channel fixed, the way you would keep a ruler.

A one-page tracker before Tet 2027

| Layer | Main number | How to get it | Cycle |
|---|---|---|---|
| Memory | Unaided recall rate (plus recognition rate) | Survey of target customers, same question every time | Quarterly |
| Behavior | Searches containing the brand name | Google Search Console, Google Trends; screenshots of Shopee and TikTok search suggestions | Monthly |
| Business | Ad cost per order, or share of orders sold at list price | Ad reports and sales reports | Monthly, compared by season |

This week, do two things: export the last three months of branded search counts from Search Console, and send a one-question survey to your target customers. That is your pre-Tet baseline. In March, measure again exactly the same way. If by then only impressions have grown and the search box is still quiet, you will know that customers saw you a lot this Tet, but did not remember much more of you.

Share this article