Availability adjustment for conjoint preference shares
At times, simulation results from conjoint analysis differ substantially from observed volumes shares in the market. This could happen for a variety of reasons, such as:
Distribution (including shelf availability): Even if a product is highly preferred to others, its availability can be limited causing its volume share to be down.
Share of mind: New products tend to have low awareness and fewer habitual customers than older products. In conjoint, because we ask respondents to review product descriptions, we expose them to old products as much as to the new ones, thus elevating awareness of new products to the degree of existing products.
Incomplete product specification: If you test only a couple of product attributes in your survey (for example, logo and price), you will not be able to represent the whole product as customers see on the shelf. Certain important features (such as weight or pack design) that are absent in conjoint may as well be the ones that drive people’s choice.
Incomplete SKU list in the test: You may be omitting some SKUs that are present in the market.
Sample definition: Your target audience in the survey is different from buyers in the market.
In order to bring preference shares simulated in Conjoint.ly closer to actual volumes shares, you can use the availability adjustment. This adjustment helps bridge this gap by assuming that all the difference is fully explained by distribution, shelf availability, or mental availability of different brands.
You may have three SKUs with the following preference shares:
|None of the above||22.82%|
However, you know that in reality volume shares are a little different:
|SKU||Actual volume share|
If you insert the volume shares from above into the availability adjustment tool, the system will suggest the following adjustment factors:
Each relative availability factor is the hypothetical percentage of retail outlets (or shelves) that stock each SKU. Generally, if an SKU performs worse in conjoint than on the market, the system will suggest a higher adjustment factor. Vice versa, if an SKU performs better in conjoint than on the market, the system will conclude that it is not as widely distributed as other products.
In this case, the suggested value of availability adjustment is 100% for Maruda and 44.08% for Kea, meaning that the former must be sold at every outlet, while the latter only in 44.08% of the stores.
Adjustments for sensitivity analysis
Adjustments of this nature can be used when you perform sensitivity analysis:
Note the difference in results without and with the adjustment:
Please note that elasticity of demand will be affected, albeit to a lesser degree than through other methods of preference share adjustment (not discussed here).
What to do with new product launches
By definition, a product that has not yet been launched will have zero distribution and awareness. Availability adjustments provide an opportunity to assess how distribution will affect volume shares.
In the example below, NPD is represented by red bars at different percentages of distribution from 0% (leftmost) to 50% (rightmost):
Points of caution
Please note that this adjustment works best:
- with small SKU counts (under a dozen) and
- when volumes shares are not extremely different from preference shares.
Published on 25 August 2019.
Read these articles next:
Dual Negative-Positive Scale
Dual negative-positive scale asks respondents to provide answers as either a positive or negative rating on a 1-5 scales. Results are calculated using the formula: Overall Score = Mean of positive questions - Mean of negative questions. View article
Subscribe to our monthly newsletter
Get the latest updates about market research, automated tools, product testing and pricing techniques.