Closing the say-do gap for more powerful insights

Background

One of the most common gripes of market research is the say-do gap: the difference between what someone ‘says’ they will do, versus what they actually ‘do’. This gives rise to many discounting research entirely, as inaccurate or unreliable.

At Vibrant, we view what people ‘say’ and ‘do’ as equally valuable, and believe in this being overlaid with behavioural economics / known principles of behaviour.

What people ‘say’ is important: it represents their views, opinions, and attitudes. Things which are critical to ask, as we can’t easily infer these. There’s a trick to asking the right questions, but that’s a story for another case study.

At the same time, what people ‘do’ has immense value. Where possible, we always recommend augmenting this with observational data. For this reason, we like to use datasets and market research panels that let us fuse survey responses with observed, real-world financial data (often for the same person), or other techniques that provide a window into behaviour. This means we can blend together what a person says, with what we know they do, based on real observed behaviour.

But how do we do this for categories where behavioural data doesn’t exist, or we’re testing new concepts?

We recently faced this provocation with our client, a much-loved consumer goods product.

In particular, they had conducted research many times in the past that predicted a favourable market result for a launch, only to find in-market results disappointing, leading to de-listing.

They needed a program which helped them get better results, particularly to inform retailer conversations. While conjoint / discrete choice modelling, or shelf-based tests can provide a more accurate and calibrated picture of potential sales, the budget and timeline did not facilitate such an advanced approach. Moreover, this research needed to inform ranging conversations with retailer, so it was important the insights were robust, actionable, reliable (and visually appealing enough to share with the retailer without re-work).

Approach

We designed a concept screening study, and conducted the research, with several lenses in mind.

1. We ensured we captured the size of the market who qualified for the study, meaning we could accurately gauge just how large the ‘addressable market’ was.
2. We captured the product suitability for certain occasions, rather than simple purchase intent overall; this helps contextualise whether the purchase is likely to be frequent or infrequent.
3. Purchase intent was captured both unpriced and priced, helping to provide a more accurate gauge of actual purchase (as pricing can significantly impact pricing decisions).
4. Results were put into context of our benchmarks, built up of nearly 100 concepts evaluated across 20,000+ consumer interviews.
5. When reporting results, we used a weighted priced purchase metric to help account for the actual behaviour of stated intent, providing a more accurate gauge of potential uptake.
6. Finally, we included testing a range of existing in-market SKUs, to help contextualise the results and better provide a comparison of in-market results.

It is important to note, of course, that behaviour doesn’t exist in a vacuum. We also noted that competitor activity, as well as physical availability in-stores (not to mention how it was stocked on shelves) can further help or hinder performance.

Result

The insights provided vital insights to inform which products to advance to market, plus further information around how best to price these.

The insights helped provide a nuanced view of portfolio decisions, helping to capitalise on various segments within the market.

Moreover, a secondary value materialised in informing optimisation of existing products in market.

And, the program has now been established, facilitating faster and more cost-effective testing of products in future, driving even greater value out of the program in future.