Response to “Publishers Need to be Data Driven”

I just read an interesting post in Book Business concerning Publishers Must Embrace Data-First Thinking.

This article has some very interesting facts, but suffers from a number of fallacies. It is a conflagration of data and facts that destroys any semblance of a conclusion. This may be rooted in the article that the blog cites, by Tom Davenport, but I would have hoped that there would have been some analysis included in the blog entry.

Data, by its nature, is the past, yet there is a blast against big publishers for making bets on books based on what succeeded in the past. A retailer (such as Amazon) can use REAL TIME data, or near real time data to adjust certain things, such as pricing and what people might be interested in from a VIRTUAL inventory. A publisher has a lag time. The data will be old, even with technology, and the time to market—even rushed—will create a lag between the recognition of the data and action on it.

Now, this does not preclude the importance of data, I am a data person from way back. But, there is data and there is information. Data tends to be raw and needs interpretation. It also tends to be badly interpreted. I can demonstrate a statistical correlation between eating tomatoes and (for instance) the number of orgasms per year a person has. Yet, the correlation is probably nothing but a statistical artifact. Still, I am sure someone would then market tomatoes as the next cure for your sex life if I came up with a study with that data.

What Amazon is doing (disclaimer, I used to work there) is not just data, but using the business concept of “spinning the flywheel.” (The concept of the flywheel effect was popularized by Jim Collins in his book “Good to Great.”) The root of this is not just data, but synergies. Because of technology, Amazon is becoming a both vertically integrated and horizontally integrated, where ever it spins the flywheel. The examples Ms. Harvey cites of positive actions by publishers are really more of the vertical integration. Combining two, or more, businesses that are in different stages of production of similar products (e.g. a farm combined with a food manufacturer combined with a supermarket). By doing this, they are indeed capturing data. But, they are more importantly in control of multiple stages and able to respond separately to each stage as they see fit. Amazon does this incredibly well and it is all part of the flywheel.

Unfortunately, for the publisher, they do not have the horizontal integration that Amazon has. This is what really spins the flywheel. While Amazon uses some data on trends and what customers wanted, it too has the issue of lag time when producing its own video, or book imprints. What it has is the ability to largely ignore the need to guess and interpret the data by letting the market (or rather its market) determine the winners and losers and then automatically the system responds. If you pay your authors (largely) by only a percentage, the authors self-select out of the system. This is not so much data, but self-correcting systems that are possible when you control a large share and are both vertically and horizontally integrated.

To Ms. Harvey’s credit, the final section of her post captures the essence of this. Workflow. When you have a vertically integrated system, with a strong flywheel, you create a workflow that always provides additional momentum to the flywheel, rather than spin against it. Amazon is fantastic at this sort of thinking and internal development. In no other company that I have observed is this flywheel effect and workflow to mesh with it better implemented. It has its flaws, including stifling innovation that goes against the flywheel, but it is massively successful in creating growth (which Amazon does for revenue and customers, if not for profits).