Study says startups can misread customer demand without coordinated tests
ESMT Berlin researchers say pricing, advertising and inventory choices must be tested together to learn how customers respond to new products.
By Lucas Ferreira · Science & Environment Writer
3 min read
Startups may draw faulty conclusions from early customer data when they change prices, ads and stock levels without treating those choices as linked, according to a new ESMT Berlin study. The findings matter for young firms and product launches because early decisions can shape both sales and what companies think they have learned about demand.
The research was conducted by Huseyin Gurkan of ESMT Berlin, N. Bora Keskin of Duke University and Rodney P. Parker of Indiana University. Their paper, “Dynamic Learning for Joint Pricing, Advertising, and Inventory Management,” has been accepted by the peer-reviewed journal Manufacturing & Service Operations Management, according to ESMT Berlin.
The researchers built a dynamic model to examine how companies make repeated choices about price, advertising and inventory while learning about customer demand over time, ESMT Berlin said. The model focuses on situations common to new products, where companies have limited information about buyers but must still decide what to charge, how much to promote and how much inventory to carry.
According to the study, those decisions affect one another more than many companies assume. Advertising can increase immediate sales while also building customer awareness over a longer period, but firms cannot directly see the level of awareness in the market. They instead observe indirect signals, such as whether customers show up or complete purchases.
That creates a difficult trade-off, the researchers said. If a company spends too little on advertising, too few customers may interact with the product for the firm to learn much about demand. If it spends too much, awareness can become so broad that it becomes harder to tell how much advertising is still influencing customer behavior.
Gurkan said, according to ESMT Berlin, that companies often expect customer behavior to reveal the right choices over time, but passive observation can point firms in the wrong direction. He said deliberate testing of both pricing and advertising can cut long-term losses and lead to better decisions as companies expand.
The study also treats inventory as part of the learning problem, rather than a back-office issue. Parker said, according to ESMT Berlin, that when a customer wants to buy and the product is out of stock, the company loses both the sale and the chance to learn from that demand. The model indicates that firms that fail to coordinate inventory with marketing tests perform worse, even when their pricing and advertising choices are otherwise sound.
The researchers found that structured experiments, including shifts in price points and advertising spending over time, help companies learn about customers more effectively. ESMT Berlin said the tests must be coordinated because pricing and advertising each affect what the company can learn from the other.
The study also found that companies do not need to keep experimenting without end. According to ESMT Berlin, a short, carefully planned testing period can help firms learn faster and improve long-term performance, reducing the disadvantage of starting with limited customer data.
ESMT Berlin said the findings are aimed especially at startups, e-commerce companies and firms in uncertain or fast-changing markets, where past customer data may be scarce and buyer behavior can be hard to predict.
This story draws on original reporting from Phys.org.