Score! The first results from The Product Recommender Engine arrived.
A group of customers received a business-as-usual email of product highlights. Some of these customers returned and bought some more stuff. Let's call that conversion rate index 100.
Another group received customized emails, recommending specific products, based on the product purchase history as decided by The Engine. These customers converted at a rate of index 157. Money was made.
Innohead is now in the process of packaging the recommender algorithm as a standard component on the platform. I wonder who else, besides booksellers, could use the system?
Innohead is picking a logo. 16 designers submitted a range of logos in the competition. From that, I and Lars picked 8.
Which logo do you prefer?
Please go to the competition on 99designs and help us choose. The 99designs voting page: http://99designs.com/logo-design/vote-uoshhc
by Innohead on Tuesday, February 08, 2011
We have our first hire, genius programmer Lars Bojsen-Møller, onboard too. Lars holds a MSc in Computer Science and Mathematics, played the piano since age 6 and loves Nick Drake, stands 192 cm tall, is pale and prefers to meet in late.
My mom sometimes asks me what the company does: "You do these logarithms, right?" Here is a rundown of the two core activities:
a) Customer Data Warehouse. We gather all the data we can find in a client company and feed it into a specialized warehouse. Call center interactions, website visits, product transactions, survey responses and so on. We use MongoDB as the database engine behind it. Love it.
b) Continuous Analytics. Based on customer data, we predict
- what leads might become valuable customers,
- which customers might be open to buy more and what they might buy,
- which customers might be at risk of exiting the company in the near future
Predictions are based on modern statistics and machine learning methods, labeled artificial intelligence in the old days. In many ways, there have been a fruitful convergence between machine learning methods and statistics, in that we are starting to understand how to analyze some of the machine learning algorithms in statistical terms.
Picking from the customers, we have
- a bank running a lead scoring algorithm
- a television provider and
- a lottery running customer exit scoring models and Customer Lifetime Value measures
- a financing and major appliances company running a product recommender
- an insurance company
- an electricity distributor
- a transportation company
Running a startup company is great fun. Does your boss suck? Do you want more freedom? Can you create real value? Then think about doing a startup.
Published with Blogger-droid v1.6.6
by Innohead on Tuesday, February 01, 2011