The short answer
MMM should account for stock availability and assortment quality using measures that reflect what customers can buy and how appealing the range is. Review stock age, size coverage, and relevant product characteristics with merchandising specialists. Use granular historical data, avoid overlapping measures, and validate the results before using them to judge media performance.
This article is part of Asked by Marketers, a series answering real questions from marketing leaders.
Sellforte's team holds more than 1,450 meetings each year with marketing leaders in Ecommerce and Retail about Marketing Mix Modeling and incrementality testing. Each week, we anonymize at least one question from those conversations and answer it in depth, based on what marketers are actually struggling with, not what keyword tools suggest. About the series →
Why marketers ask this
A campaign can bring shoppers to a product page and still generate few purchases if the sizes they need are missing. Sales can also rise when a retailer receives a stronger range, even without a better campaign. The marketing team needs to understand those changes before deciding to cut or increase spend.
Assortment also changes the offer a campaign promotes. The same number of emails can produce very different sales when one email features fresh, appealing products at attractive prices and the next has a weaker offer. MMM needs enough information about the products to help explain that difference.
Which stock and assortment measures should MMM use?
Choose measures that describe the buying constraints and product characteristics that matter in your business. Availability describes whether customers can buy what they want. Assortment quality includes the appeal of the range, such as relevant brands and fresh products. A stock total alone cannot describe all of those conditions.
In Marketing Mix Modeling, these commercial factors help explain sales alongside media activity. A fashion retailer may need to understand missing sizes and stock age. A marketplace advertiser may first need to separate whether a product is available from the additional visibility purchased through sponsored ads.
On a small screen, scroll sideways to read the full table.
| Business question | Possible measure | What to check |
|---|---|---|
| Can customers buy the sizes they need? | Size coverage within the range | Establish whether the data describes warehouse stock or the products currently offered to customers. |
| How fresh is the stock? | Stock age or agreed age groups | Confirm what the age field means and whether it is available at product level. |
| Has the type of stock changed? | Stock source or acquisition category | Check whether this adds information beyond stock age and discounts. |
| Does the range contain products customers want? | Presence of relevant brands or product groups | Review the definition with merchandising. Stock age and source may leave important differences in product appeal unexplained. |
These are candidates to assess, not a mandatory list. A stock-source field can be informative when different sources bring different types of products into the range. In another business, that distinction may add little. Involve the people who understand buying, merchandising, and sales forecasting when selecting the measures.
Keep pricing and discounts alongside the stock information. Fresher or more appealing products may also have different discounts. The modeler needs to understand the offer as well as the product characteristics.
How should assortment information become a model input?
Give the modeling team the underlying product information and agree how to turn it into useful measures. A single monthly assortment score may be too coarse to explain changes in a model using daily data.
Where available, product-level sales records can carry fields such as stock age, stock source, and size coverage. For example, a record can show how long an item had been in stock when it sold. That gives the modeler more detail than an average stock-age figure for the whole business.
Check that each proposed measure changes enough over the modeled period to be useful. A commercially important characteristic can still be difficult to assess if it barely changes in the available history. Check how informative the measure is in each market.
Then examine overlap. Stock source and stock age may describe much of the same change in the range. Including several versions of the same signal can make the results difficult to interpret. The modeler may combine source and age into agreed groups, or retain the measure that best represents the business situation.
Document how the range is supplied and who makes those decisions. A stock mix set outside the marketing team has a different business explanation from a campaign choice. The modeling team should understand that process before deciding how to use the variable.
How should marketers use the assortment results in budget decisions?
Use assortment information to judge marketing performance in its commercial context. A stronger range can support higher sales, and a weaker offer can make a campaign look less effective. Review the media activity together with the offer it promoted when discussing the budget.
Be clear about which decisions the marketing team controls. If its remit is media and CRM, assortment analysis helps explain the sales result and informs discussions with merchandising. It does not make the marketing team responsible for buying decisions or turn a media recommendation into a recommendation about the range.
When sales fall below plan, use the model to assess what a change in media investment could contribute. Review that scenario alongside the commercial team's assessment of the offer and the business's profitability goals. A large assortment contribution alone does not determine whether the next media investment is worthwhile.
How this looks in practice
Consider a hypothetical fashion retailer that sends a promotional email each morning. The audience size is similar on two days, but attributed sales differ substantially. One email features fresh stock from popular brands at attractive prices; the other promotes an older range with fewer sizes available.
The team brings stock age, source, size coverage, and discount information into the data review. It checks which fields are available at product level and whether stock source and age describe the same differences. The modeler agrees a manageable set of measures with the merchandising specialist.
MMM then evaluates the marketing activity alongside those commercial factors. The difference in attributed sales alone does not establish how much incremental revenue either email generated. The team reviews the model's media estimates and uses relevant experimental evidence where available.
Before changing the email plan, marketing and merchandising review which offers drove the difference and whether the chosen assortment measures describe them adequately. That gives them a more useful discussion than comparing send volumes alone.
How Sellforte helps
Sellforte models media alongside commercial factors affecting retail sales, including assortment characteristics where suitable data is available. The team selects usable measures and checks how they affect the interpretation of marketing performance. Book a demo to discuss how your stock and assortment data could support MMM.
Authors
Lauri Potka is the Chief Operating Officer at Sellforte and has over 15 years of experience in Marketing Mix Modeling, marketing measurement, and media spend optimization. Before joining Sellforte, he worked as a management consultant at the Boston Consulting Group, advising some of the world's largest advertisers on data-driven marketing optimization. Follow Lauri on LinkedIn, where he is one of the leading voices in MMM and marketing measurement.