Insights

Is purchase intent a reliable predictor of sales?

Dr. Outi Somervuori

Sep 22, 2021

When doing consumer research, brands often wonder how they should interpret concept test results. What does it mean when purchase intent for a food product scores 65 percent? Does the score reflect the real purchasing behaviour? Is the score reliable even though the respondent did not taste the food product? Research shows that purchase intent and recommendation responses are actually the most accurate predictors of real food purchase in a consumer research setting.

In Elina Kytö’s doctoral dissertation (2020), she describes how effectively different consumer research measures predict real purchases in the food sector. In her tests, she used the same purchase intention measure that is utilized in Cambri: 5-point Likert scale question of willingness to buy where 5 indicates a high probability and 1 indicates low probability. The test set-up included three phases:
1. A preliminary test to examine the respondent's expectations towards the product
2. Tasting (for respondents on the tasting condition)
3. A few weeks after the research participants in both conditions reported which products they had actually bought.

In total, 467 respondents participated in the test, and 346 respondents completed all 3 tests.

How effectively does purchase intent predict real purchases?

Companies are often very well aware of the fact that marketing efforts and competition influence purchases. However, when doing consumer research, they can easily discount the results of purchase intent questions as irrelevant or unreliable. However, this may be a flawed interpretation. Purchase intent questions can offer valuable consumer insights.

For example, Ramanujam & Tacke (2016) suggest that there's a 50 percent likelihood that the respondents who indicate a high probability to buy the product (choose number 5 on the scale) will actually purchase the product. For those who choose number four, the probability is 10–20 percent. While these general rules of thumb may sometimes be very helpful, their accuracy depends for example on, e.g. from whom you ask the purchase intent and on what kind of information you provide for the respondents.

Consumer research offers more accurate results when conducted among a relevant target market

Kytö’s test showed that high purchase intent likely leads to real purchase if the respondents are users of a brand or heavy users of a category. These results highlight the importance that companies need to segment their markets, select attractive target markets, and conduct their concept test also among the attractive target market to get useful consumer insights.

Purchase intent is easier to predict when more product information is available

It might seem risky to provide the respondent with a lot of information when doing consumer research. However, Kytö's research shows that it might be a good idea to actually provide the respondent with more information rather than less. In Kytö’s case, the results proved to be more accurate in conditions where the respondents were told the brand of the product in comparison to test conditions where the respondents evaluated a brandless product.

Tasting is not a prerequisite for an accurate result

The test included two conditions. In the first one, the respondents rated their purchase intention based on brand and package with the help of a picture from a normal buying situation. In the second, the respondents rated the purchase intention after tasting the product. They also had the brand and package information at their disposal. In both cases, the purchase intent was an equally effective predictor of real purchase.

What does this mean?

These research results suggest that companies should keep measuring purchase intent before making launch-related decisions.
The results also suggest that the concept test KPIs available in Cambri are reliable measures of actual purchase.
It is highly recommended that companies should segment the markets, select attractive target markets and conduct the concept test among those attractive target markets when gathering consumer insights.
It is a good idea to include all relevant information, such as brand name, in the concept testing phase to improve the accuracy of the results and learn from the consumer insights.

Sources:
Kytö, E., 2020, From purchase intention to purchase behavior.
Ramanujam & Tacke, 2016, Monetizing Innovation: How smart companies design the product around the price