Aventine Partners

Limits to Data

Via the Freakonomics Blog I ran upon the departure announcement from Doug Bowman,  the head of the Google visual design team.  What is one of the driving forces behind his leaving Google?  Too much reliance on data to make decisions.

Here at Aventine we naturally highly value data.  But we also often talk about the limits of analytics as a science, which is to say experience and even instinct bring much to the process.  Typically experience makes the difference not so much between right and wrong or more or less accurate, but between more and less useful.

However, the Bowman post prompted some new questions, though he doesn't frame them this way much less answer them.  Is the problem with too much data and analysis:

1. It doesn't always lead to the "right" answer?
2. It is dehumanizing?
3. It is a nuisance?

My instinct (if that's allowed) is that it's a bit of all three, but especially one and two.  There is much that can't be measured and reduced to formulae, and part of our dignity as human beings lies in the fact that we do not mechanistically respond to events or stimuli like inert matter or mere animials. 

This is very relevant in the realm of predicting behavior, as we often do at Aventine, but even more so when dealing aesthtic questions and some the cognitive processes that good information visualization strives to enable.  In a funny way the power of our models keeps us employed, and the magnitude of the error terms keeps us human.

Topgrading Marketing Spend

As I noted in my post To Cut or Not to Cut?, this is a smart time for most companies to reduce marketing spend, notwithstanding desperate pleas from marketing services providers that you double-down in the downturn.

Is there a smart, systematic way to do this?  Probably, but it's elusive, as confirmed by the many stories I hear of the across-the-board, arbitrary 10%, 15%, 25% cuts many marketers are trying to cope with.

A recent white paper from former colleagues at Deloitte Consulting prompted my writing this post, so let's look at their approach, which they call Structural Cost Cutting.  

The premise of the Deloitte's Structural Cost Cutting is that the big opportunities for big near-term reductions in spend come from challenging basic assumptions about the business and how marketing supports it.  The savings will emerge from multiple categories of marketing spend.  The Deloitte hit list of areas to examine are:

More on R and Alternative Statistical Analysis Packages

As a follow up to my prior post on the R statistical analysis tool, there have been a couple of interesting recent blog posts on the same topic, notably:

Comparison of data analysis packages: R, Mathlab, SciPy, Excel, SAS, SPSS, Stat, from from the AI and Social Science Blog, and

How Google and Facebook are Using R from the Data Evolution Blog.

Some quick takeaways:

• Google and Facebook are big users of R, which is the most popular analytics tool at Google.
• They use R for development, but not implementation of models, for which Python or C++ are the tools of choice. 
• The biggest debate seems to rage between users of R and SAS.  Many, many interesting takes on this question if the comments to the first post. 
• Ability to handle large data sets is one of the big differentiators and drivers of preferences.
• Some contend there is an age split, with younger users tending to prefer R and older users tending to prefer SAS.    

In any case, these make interesting reading for those facing these decisions.  Enjoy. 

To Cut or Not to Cut?

As has been in the case during every downturn I've lived through as a professional (this is my third), many observers of the marketing discipline are quick to remind us that "studies show" successful companies increase advertising spend during downturns and reap great rewards from having done so during recoveries.  Of course, those most emphatic in offering this counsel are those peddling media and marketing services.

One of many possible examples can be found in a September blog post by John Quelch offering advice on "How to Market in a Recession".  Quelch is a director at WPP, so there may be some bias when he counsels: "Maintain marketing spending.  This is not the time to cut advertising.  It is well documented that brands that increase advertising during a recession, when competitors are cutting back, can improve market share and return on investment at lower cost than during good economic times."

Even though we also peddle marketing services, I question the wisdom of this line of thinking.  More to the point, we need to be mindful that what was true in the past may not longer hold.

What's different now?