November 13, 2014

Competing With Zulily

A fellow consultant called me - asking why catalogers are so "pig-headed" when it comes to testing? This individual was frustrated, because the catalogers the person works with will not test pagination, will not execute holdout tests, will not serve up different versions of the website to different customers, will not test personalization, simply "will not" experiment.

"Why is this the case" ... this person asked? Maybe you can answer the question.

Contrast the point of view above with what Zulily is trying to accomplish ... a different version of the website for every single customer (click here).

Now, the conclusions in the article are completely bogus (tenure on the website drives increased purchases - no - the best customers do everything, and have been around the longest, and therefore, they evolve to high-frequency segments much faster than everybody else - and this happens at companies who never change the website too, and maybe the firm helping Zulily wanted that tidbit put in the article for unknown reasons, that's a topic for another day). But the concept of testing, of trying things, of doing something different for every customer, that's going to pay off, and is valid to discuss.

The marketing world is moving in several different directions, simultaneously.
  1. Catalogers are largely "doubling down" - focusing on the 60-75 year old core customer. Here, you do not need to be a sophisticated marketer and you don't need to compete with Zulily - you simply need to offer products that cater to a 60-75 year old customer. Lots of short-term profit to be had here, with undeniable long-term consequences.
  2. E-commerce businesses are heading down the Zulily path - an algorithmic version of marketing. Just ask Amazon. You don't do retargeting because it works, you do it because of the data you collect. E-commerce and Artificial Intelligence are going to become the same thing. Here's the danger - algorithms work really well on commodity items at low prices. Algorithms struggle with fashion-centric and excitement-centric business models.
  3. Retailers will not focus as heavily on algorithmic solutions - they will lag behind e-commerce folks because of all the compromises they have to make to drive customers into stores. Retailers will have bigger issues - in-store fixed costs are going to demand attention. But retailers will be pulled into tech AI solutions ... there's a reason why Google and Macy's tout online search that drives in-store purchases ... Google wants the searches for $$, they want the mobile searches and online visits to fuel their AI engine. Retailers will be pawns, offline inputs for data-starved AI engines (Google, Facebook, Twitter, IBM, Microsoft, and many others).
  4. Mobile will evolve differently than e-commerce - much more of an AI / algorithmic solution that yields convenience, one that yields hyper-personalization among folks under the age of 35. Just use Google Maps for travel, and you'll see where all this is headed.

Page Counts When Bifurcation Helps Best Customers

Yes, your catalog customer file bifurcated ... with a minority of best customers needing more catalogs and a majority of your total file ...