Sur La Table Q&A: 11 Ways To Ruin Your Data Analysis

Sur-La-TablePulling insights out of disparate data sets is a challenge for many companies. Sometimes the problem is a marketer’s approach to data analysis. Kevin Ertell, e-commerce VP at kitchenware retailer Sur La Table, identified at Experian Marketing Services’ Client Summit 11 ways marketers undermine their efforts. AdExchanger spoke with Ertell about those points and other best-practice tips.

AdExchanger: How do you analyze your data?

KEVIN ERTELL: We’re looking at trends to understand where things are heading, but we have lots of data coming from lots of systems and it doesn’t always match. So the analysis we put together needs to have four attributes. It needs to be focused, it needs to be actionable, it needs to be manageable and it needs to be enlightening. That’s what I call FAME.

I like to spend time thinking about probabilities. I look at ROIs and I’ve built this spreadsheet that does a Monte Carlo simulation that allows me to run lots of different possibilities and feed a bunch of variables into it to predict lots of different possible outcomes.

Can you give me a specific example of how you do that?

Every year we put together a budget and look at where we want to spend capital dollars to make improvements to our website. Right now we’re just about to add PayPal to the site. So what we want to do when we look at adding PayPal is look around the industry to understand what types of benefits people are seeing from PayPal.

We want to understand what types of customers they [other retailers] have, take a guess on how many of those people would use PayPal on our site, what their average spend might be in comparison to our average normal spend, what type of increase in conversion we might get because of that next payment method, etc. Then we look at it on a desktop versus a tablet and a smartphone. There’s a lot of data out there saying 60% or more of smartphone transactions are paid for on PayPal. That’s a huge deal. We had to factor in all of that to make a guess at what kind of return we would get.

We didn’t look for a specific a number, we did it as a range of numbers and based our investment on the low end of that range. What we want to know is if we make an investment, are we going to get a return at x period of time, even if it performs at the lowest range?

Would you say that you’re your company’s chief data scientist?
We don’t have anything like that. I’m not an analyst, I’m a decision-maker. We do have a variety of people doing analyses but I’m mainly talking about our e-commerce team.

What are you focusing on over the coming months? What’s on your roadmap?

We’re continuing to enhance the functionality of the site; we’re about to add PayPal and we are looking for more ways to deal with the explosion in mobile traffic. It was growing quickly last year and it’s so much faster this year. We’re definitely spending time on that front to make sure we have the right experience, whether someone is shopping on our site with their mobile device or in our stores. The open rates for emails on mobile devices are also huge and I don’t see that getting smaller anytime soon.

Kevin Ertell: 11 Ways People Kill Good Analysis

  1. Hiring reporters instead of analysts to do the job. Analysts must be creative in order to find patterns in data that lead to insights.
  2.  Turning analysts into reporters. Asking analysts for all sorts of data can be a distraction. Ask focused questions instead.
  3.  Expecting the data and the analysis to be flawless.
  4.  Failing to define objectives.
  5.  Asking for numbers just for numbers’ sake. “There is an old saying that you can’t manage what you can’t measure,” Ertell said. “You also can’t manage everything you can measure.”
  6.  Insisting on simplicity over accuracy.
  7.  Looking for a hard number. Marketers “need to get comfortable with ranges,” Ertell noted. “We need to know there are other possibilities and plan appropriately.”
  8.  Lacking a solid understanding of statistics.
  9.  Expecting answers immediately.
  10.  Ignoring your gut instincts.
  11.  Presenting the data poorly. “If you have more than 2 data points, don’t use a pie chart.”

Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!