Brand Lift in Tajikistan: How to Measure Brand Awareness Growth

Clicks and impressions tell you how many people saw an ad, but they say nothing about the main thing: whether anything has changed in the audience’s minds. Do people know the brand better now, are they considering a purchase, did they remember the message? A Brand Lift study answers these questions. Below is how to run one in Tajikistan and turn the results into media plan decisions.

How the study works

Two groups are compared. The test group was exposed to the campaign; the control group did not come across the advertising. Both are asked the same questions, and the difference in their answers shows the effect.

Four indicators are usually measured:

• Ad Recall — whether the person remembers recently seeing advertising for the brand.
• Brand Awareness — whether they are familiar with the brand.
• Consideration / Preference — whether they are ready to choose the brand when buying.
• Message Association — whether they link the campaign’s key idea to the brand.

Lift formulas

in percentage points: Lift = Test − Control
in per cent: Lift = (Test − Control) / Control × 100%

Illustration: if 42% of respondents in the test group know the brand and 35% in the control group do, the lift is 7 pp, or 20% relative to the control level.

How to design the test and control groups

Step 1. Random split

The audience is divided at random: one part is shown the campaign, the other is not. For outdoor advertising and local campaigns, geographic pairs work well — districts or cities with a similar population make-up.

Step 2. A short questionnaire

1. Category screening: whether the respondent has bought such products in the last six months.
2. Unaided awareness: which brands in the category they can recall on their own.
3. Aided awareness: whether they have heard of the named brand.
4. Ad recall: whether they have recently come across advertising for the brand.
5. Intent: whether they are considering a purchase in the next month.

The longer the survey, the sloppier the answers. Keep the questionnaire to a minute or a minute and a half.

Step 3. Sample size

The margin of error for the difference between two proportions can be estimated as follows: 1.96 × √(p₁(1−p₁)/n₁ + p₂(1−p₂)/n₂), where p is the proportion in a group and n is the number of interviews.

Sample calculation: with an awareness level of about 35% and a thousand interviews in each group, the margin of error for the difference is roughly ±4 pp. A smaller lift cannot be reliably detected with that sample. If you expect an effect of two or three points, increase the sample several times over or extend the campaign. A survey of a few hundred people will only show the direction.

Step 4. Data quality

Add attention-check questions, set a minimum completion time, and verify region and demographics. Discard questionnaires from people who “recognised” every brand in a row within seconds.

When you can trust the effect

1. Confidence interval of the difference (95%). If it includes zero, it is too early to talk about lift.
2. Significance. For two proportions, use a z-test or chi-square, with a threshold of p ≤ 0.05.
3. Segments. Sometimes there is overall growth but none in the core audience. That is a reason for an additional test.

A conclusion in the report might read like this (figures are hypothetical): “Awareness rose by 6 pp, the result is significant; among 18–34-year-olds the lift is above average.” Or: “Purchase intent did not change — we are revising the offer and extending the test by two weeks.”

How to translate the results into a media plan

1. Messages. Keep the wording that strengthened the link between the brand and the key idea; drop the weak lines.
2. Frequency and length. If the advertising is poorly recalled, increase the number of contacts and try 6- and 15-second spots.
3. Channels. Add budget where the lift is higher — for example, to video if it outperformed banners.
4. Funnel. Awareness went up but purchase interest didn’t? Add guarantees and reviews to your landing pages.
5. Regularity. Repeat a shortened measurement every 6–8 weeks.

An end-to-end example (figures are hypothetical)

• Objective: raise brand awareness among Dushanbe residents aged 18–34.
• Media: video plus outdoor billboards, four weeks.
• Survey: 1,000 interviews in each group.
• Result: awareness in the test group 41%, in the control group 34%, a difference of 7 pp, which is significant (p < 0.01).
• Decision: expand video, keep the successful opening of the spot, strengthen the landing page, repeat the measurement in six weeks.

Mistakes that undermine the measurement

• An overloaded questionnaire: answers become random and fieldwork gets more expensive.
• “Before and after” without a control: seasonality and news distort the picture, and the contribution of advertising can’t be isolated.
• Mixed segments: the effect gets diluted; analyse the core audience separately.
• A small sample: a couple of hundred questionnaires won’t reveal a shift of a few points.
• Detachment from sales: Brand Lift complements business metrics rather than replacing them.

Pre-launch checklist

• The main metric has been chosen: awareness, recall or purchase intent.
• The method for forming the groups is clear.
• The questionnaire has no more than five substantive questions.
• The sample size has been calculated for the expected lift.
• There are rules for cleaning data and stopping fieldwork.
• Hypotheses and a plan for A/B testing the creatives are ready.
• A follow-up measurement has been scheduled.

If you want to assess how a campaign affects brand awareness, SKY Central Asia can help plan the study and run the survey — find out more in the section on media research. To build the findings into your placements, see our media strategy service and the strategy and research area.