AI industry discussions VLOG /

#001 - Consumer Research

April 20/45 min read
  • Björn Treje
    Björn TrejeCustomer success manager
  • Peder Ribbing
    Peder RibbingHead of expert services
  • Alexis Bolonassos
    Alexis BolonassosProduct owner at Nepa

In our pilot episode we are joined by Alexis Bolonassos. He is a product owner at Nepa, an exciting global consumer research company. We scrape a little on the surface to the question about how AI is shaping the future of consumer research? What are the key opportunities for AI in consumer research, and what are the challenges?

(The discussion is in Swedish)


Key takeaways from this episode

Alexis talks about one of Nepa's retail customers where they are using AI in several interesting ways.
  • Instead of centralized analytics in one decision making role or department, we decentralise it and put the power of AI in the hands of the local entrepreneurs, the local store owners. 
  • We get insights for purchase behaviour in certain stores and compare on national levels - to identify sales increase opportunities for individual stores.
  • Several hundred stores continuously get customised recommendations and insights so that they can optimise sales. 
  • Also uses purchase behaviour in order to match the most relevant surveys to member database - in order to get better insights.
  • By combining different data it is possible to  create many new business valuable predictions. 

An area where a lot is happening right now is media mix modelling.
  • Investments vs results, in this case bought media vs sales.
  • Using technique to isolate effects, remove rainy days and other event that otherwise would be bundled in with the end result.
  • In order to identify the best possible media mix for sales.

It takes courage to invest in predictive models.
  • It’s an iterative process
  • Start small, but measurable
  • Be prepared that not all experiments will succeed. It’s part of the process.

Many have huge expectations on AI due to all the hype.
  • The concrete applications can create more value but may be less exciting than initially thought. 
  • More specific projects around efficiency creation than big general solutions. 

Assessing the quality of survey data can be hard


Many survey responses are bots
  • It’s getting increasingly harder to discern between bots and humans.  

Data can never say what a customer wants by itself.
  • We need to understand WHY they want something. 

Data is often used as an excuse NOT innovate.  
  • Risk aversion is common.
  • Data and research often used as a safety blanket to avoid responsibility.
  • But to win with data you need the courage to act on what the data doesn’t spell out clearly.

Technology in itself is not a differentiator.
  • It is the people and the way they use they use the technology