Case Study: Analytics Teams

Screen Shot 2015-09-04 at 10.03.50

(This is a business case study. It will be used to guide discussions during the session: “Analytical Teams” at the Vendo Partner Conference in Barcelona on Wednesday, September 14th.)


Thierry was supposed to be on vacation. He’d spent the day skiing with his family in the Pyrenees. It was Tuesday night at 02:15 in the morning. Or Wednesday morning now that he thought about it. He was looking at the day’s numbers and he was growing more annoyed by the minute. The numbers on his screens were blurry. It wasn’t just the late hour. He rubbed his eyes. Then he took a break. Five minutes later he went back to his screens. He saw them more clearly now but the numbers still weren’t making sense. He had been trying for hours to get to the reality below the figures. He wanted to find a root cause that would explain the downward shift he was watching.

“Don’t think I’m going to be sleeping much tonight,” he whispered to himself quietly. He was very quiet, sitting still in the night. It would have been satisfying to let out a loud “¡Joder!” but that would have woken up his kids. He didn’t need to add to his problems.

Earlier in the day he had driven by a frozen mountain lake on the way to the ski resort. He pulled over, parked and walked up to the edge of the lake. He could see life underneath the ice. His kids were fascinated. They stood there for a minute watching the aquatic plants move with an invisible current.

The insights he was looking for were also just below the surface, fuzzy and unclear, like they were under a layer of ice. He felt like he was very close to figuring it out. But he wasn’t sure. How could he crack the surface to get the insights he needed?

This wasn’t the first time he’d asked this question. And he felt certain that it wouldn’t be the last. He reflected on how he had gained new insights as his company had grown. How had he broken through? He began to think about the people on his analytics team.

“Who should be by your side in a situation like this?” he asked himself. He recalled that as his company grew the answer was always different. It depended entirely on the situation. Each time Thierry needed a new member for his analytics team he found himself looking for a new type of analyst.

He thought about the history of his company. He also thought about the experience of his partners’ companies. There were clear patterns. A company’s analytics needs change based on their size. He wrote down his thoughts:


You start small. You and a buddy have a great idea. You are working out of your apartment with frequent trips to Starbucks for caffeination and a change of scene. The analytics team is you and / or your partner. The primary tools will be free ones like Google Analytics and whatever reporting comes from the companies with which you are buying and selling. You are focused on key metrics like how many sales you make per day.

For the smaller team your new analyst doesn’t need to be full time. He or she can have other tasks, too. He or she will point out big changes (ex. a drop in sales) and research the causes. You’ll use this information to focus your efforts to fix or improve areas of your business.


As your company grows – using insights from your current analytical tools – you will want to expand your analytical power. As a medium sized company you can now justify the cost of a dedicated analyst. You could also shift your part time analyst to full time but you may need a different person at this point.

You’ll be looking for someone to expand the number and breadth of tools you use. You will be paying for licenses for these tools, dedicating servers to store the data and operate the tools. You’ll be working to make your data “clean” and useful for your business. Increasingly analytics will be involved in key decision making. You will be using your analytics to find out “what is happening?” and developing ideas of “what could happen?” You will be generating new ideas by looking into the data and testing hypotheses.


As you grow larger – thanks to hard work, luck and growing analytical power – you’ll be looking to hire a leader who can expand your analytical team. You will be hiring someone to take you beyond your current analytical tools into automated decision making and action. At this point you will be focused on artificial intelligence that can learn to optimize your key processes. You are taking humans out of the Data > Analysis > Decision > Action (DADA) process. You will be focused on getting the right data into the system and measuring the results, then tweaking and improving. There will still be areas where you do traditional analytics but the majority of your efforts – and therefore your analytics team – will be focused on artificial intelligence.

In summary, the size and composition of your team depends entirely on your company’s situation. A small team needs someone, usually part time, who is working with existing tools to provide insights. A medium sized company will have a dedicated analyst and growing influence on how the business functions. In the larger company the analytical role shifts toward automating decision making and taking action. Who you need by your side depends on where you are.

Thierry shared his thoughts and the data on his screen with his analytics team. It was getting late. He knew that when he woke up his team would have fresh insights and perspectives for him. It felt great to have them by his side. He shut down his computer and went quietly back to bed.

Questions for discussion:

  • Who do I need on my analytics team?
  • What skills do they need?
  • What should they know when I hire them?
  • What do I need to teach them about my business so they can be effective?


A note on hiring:

One of the challenges of hiring new people is that you often have to fire them quickly. It’s rare to find cases where the hiring success rate is above 50%. Fortunately, analytics is a field where you can find out someone’s capabilities by giving them homework before they start. As part of the hiring process you can ask them to do real work that is very similar to the work they will do on the job. Whereas Google used to ask potential hires silly questions like, “How many golf balls fit in a 747 passenger plane?” you can ask a job candidate to review your data and provide actual insights.

Leave a Reply

Your email address will not be published. Required fields are marked *