Thursday, November 12, 2015


Les tendances de la Valley, par 1000mercis... par PetitShow

I wanted to write this blog post after having watched this video. CEO of 1000 mercis Yseulys Coste is speaking about the hypest words in the Silicon Valley at the moment: Incrementality and people based marketing. I wanted to get some thoughts on the first item.

Indeed, Incrementality is the graal of marketers: Being able to have a clear identifiable impact on sales, and hence, on the results of a company. Today, thanks to big data, we have more and more information that could help in this quest of identification. Same thing for category management and retailing: If I enter a new SKU, how much extra sales could I expect from it? Moreover, how much it will cost me to extend my product range (the infamous COGS I discussed several times here).

But this quest is far from being easy. Nowadays, we have so many factors that could impact sales that it is difficult in a 360 marketing approach to measure individually each factors:

  • The quality of the product: Something that people tend to forget, but if you have an average innovation launched in a crowded market, your elasticity will be low
  • The mass media advertizement. Even though going in limbo compared to Internet and social media budgets, mass media like TV, radios or billboarding are still very efficient sales booster, even though it is hard to link direct sales to it.
  • Retailing effort: You know it is my forté, but you need to invest in your retailing channel (as a brand). End of the aisle display, interractions with the shopper, make sure you have the right stock available in store not to miss sales.
  • Of course, Direct marketing campaigns, including social media, emailing, or even website visiting, where you can have a lot of info about the shopper path toward the sales.

Here are also some aspects you need to take into consideration:

Behavioral Data
One of the key component of segmentation today is to be based on customers behaviors. Indeed, social and geographical segmentation is no longer applicable, as customers adopt their consumption based on what they believe rather than historical background.

Dividing your Data Base into 2 Groups
One of the best ways to measure incrementality is to have two groups of people: One who will see the ad, and another one that won't see it (for example, one will receive an email, the other don't). Three aspects to keep in mind:

  • Let's not forget that you won't master the other aspect of your 360 marketing effort. For example, do you know that your first group of people has seen no products in its local store? It has not seen the billboard campaign? It has been away on vacation at the moment? 
  • One of the issue that I have is that it is ok to test, to experiment, in a R&D effort. But if you really want to have a deep impact on sales, it is also important to have an impact on 100% of your data base. And that is something sometimes CRM experts tend to forget.
  • You need to have large enough groups, and specifics group, in order to extract real information of the testing. If you do it on a data base lower than 10% of your overal data base, the info won't mean much. Moreover, if you have only 100 people for the test, it won't really be good infos either.

The topic is wide and interesting. I 'll probably write more in posts to come.