Analytics Team Dissolution – Unique Case Study

This unique case helps to understand how to build and lead an effective business analytics team, by learning from the sharp decline of one such team.

By Randy Bartlett, Sept 2014.

To understand how to build and lead an effective business analytics (BA) team, we studied the sharp decline of one. We spoke with several people involved, before and after a change in leadership.

Let's start with a synopsis.
  • Company's industry: Short history of analytics on commercial side; dominant sales culture
  • Company: $10 billion plus, midsized global
  • Business need: Key business unit underperforming
  • BA location: Inside commercial operations
  • BA composition: About 50 professionals (10 quants, 30 light business analysts, 10 other); some light analytics at a comparably sized sister group
  • Leadership: BA was founded under a seasoned business professional acting in the capacity of a chief analytics officer (CAO). This original CAO reported one level removed from the CEO and was well supported, but the C-level did not appreciate the value of business analytics. The CAO had six reports, including two strong lieutenants managing the quant functions. As a result, there was enough domain knowledge and statistics to provide capabilities above the industry norm.
  • BA purview: Marketing research, predictive modeling, forecasting, et al.
  • BA clients: Four business units managing products (65-70 clients); 15 fact-based clients, remainder were opinion-based clients

Analytics Team

Now, let's take a look at what happened.

The original CAO built and managed the practice using a delegation-style approach, which depends on hiring, trusting, and supporting smart people. The CAO, who possessed a strong business acumen and had great relationships, acted with integrity by placing the company's interests first. The CAO had light training in statistics but no data analysis experience.

The CAO chose lieutenants who had the courage to report the facts and viewed the BA group as objective stewards of information. The lieutenants needed a supportive CAO to help navigate any mercurial opinion-based clients.

In the fullness of time, this CAO moved on, and the BA group merged with its sister group. The "replacement CAO" -- an experienced general business manager with no BA experience -- insisted on reviewing all findings before passing them on to the business units. This CAO insisted that findings matched business opinions and projected desired perceptions; anything inconsistent was discarded. Because the CAO didn't support lieutenants as they dealt with opinion-based clients, these clients felt free to push harder to make results fit their opinions and to make their performance look good.

One business unit continued to underperform, but objective information that would help guide performance wasn't available any longer. Nevertheless, all the opinion-based clients looked like high performers.

The following table summarizes the contrast between the two CAOs:

Table 1:
Original CAO Replacement CAO
Employed delegation-style approach Employed command & control approach
Accepted new facts Managed perception
Valued BA for the information Valued BA for managing perception
Covered lieutenants when making waves Did not support lieutenants making waves
Valued stewards of information Valued yes-men

In the new environment, the lieutenants became so undermined that they were irrelevant. As one said, "Whatever you analyzed, they [the opinion-based clients] were going to do what they wanted anyway."

This gap provided opportunities for yes-men to serve up nonsense numbers to fit conclusions by hiring outside name-brand management consultants. The yes-men were too naïve or disinterested to realize the analysis they were buying at a premium price was dubious. For their part, the management consultants were happy to supply whatever findings would fulfill the desired conclusions… and their sales quotas.

The yes-men began outsourcing everything, whereas the two lieutenants had believed in having some native BA capabilities. At times, the management consultants failed in an obvious way, and internal resources had to be repurposed to rescue them. Meanwhile, the fact-based decision makers became wary enough of the CAO's spin that they hired their own outside consultants.

Eventually, the lieutenants left, and BA stopped providing the services it was expected to provide. All that remained from the BA group was the dubious external consulting, which was filtered through the replacement CAO. This CAO wasted a pile of money on the dubious analysis and later on a big-data fishing expedition.

There are four observations to draw from this example.

  1. Corporations must be willing to get their hands dirty. They need an empowered CAO with as much BA experience as possible to lead. Similarly, the C-level must be willing to remove number spinners. When the original CAO left, the obvious move was to bring in someone with CAO experience or promote one of the lieutenants.
  2. A delegation-style approach hinges on hiring right. Table stakes for quants include two years of training in statistics/data analysis, along with industry knowledge or help from industry experts.
  3. A well-supported review process could have maintained the integrity of the numbers.
  4. When outsourcing, hire the training, not the brand name. If you have internal quants to manage the external ones, all the better.

Do you take away any other lessons from the dissolution of this analytics group? How else might the BA group have survived while serving a useful purpose? Are you surprised that this sort of situation happens within companies today? Why or why not? Share your opinions below.

The Dissolution of an Analytics Team

Randy Bartlett is a Business Analytics Author & Thought Leader.
His latest All Analytics blog:
Analytics Magazine:; Amstat News: Workshops: