My friend James Connolly is a legendary tech writer. I had the privilege of working with him back in my days at the Boston Herald. He’s been writing about technology longer than I’ve been writing about measurement!
Back then at the Herald, Jim, like all of us, spent his days translating complex concepts into eighth-grade English. It is a skill that has served him and his readers well. He is now editor of All Analytics, a great newsletter for us data geeks, as well as UBM’s InformationWeek. (See more great newsletters for data fans at “The Joy of Data: The Best Newsletters and Podcasts for Data Fans.”)
Here are two sets of quotations from his work at All Analytics that capture the wisdom of decades of analytics experience. The first is on life as a data analyst, the second on getting buy-in for a program. And if you want more of Jim’s wisdom, links to all his posts are here.
Words for data analysts to live by
The following quotes come from a SAS Analytics conference panel that Jim covered. It featured Daniel Thorpe, vice president for enterprise analytics at Lowe’s Companies Inc., and Ed See, principle for customer transformation at Deloitte. The title is what caught my eye: “15 Learnings That Take Analytics Professionals Years to Figure Out (or Rather, Had I known These When I Started My Career, I’d Retire Twice as Rich).” Interestingly, according to Connolly, “Much of the discussion centered not on data itself, but what analysts and their employers do with that data, relationships, and what happens when things go wrong.”
- “If you don’t know what success is, it cannot be measured.”
See explained that analysts want to know not just the business goal of a project, but:
- what questions must be asked,
- what metrics are going to be used to measure success,
- what data will be needed,
- what level of quality is good enough for the project, and
- when the business needs answers.
- “Variation is our friend.”
Thorpe said he “learned the hard way” that variations in data shouldn’t be an aggravation but should be something that analysts seek out, that they can learn from when some data just doesn’t make sense.
- “Do not apologize that your analysis is not what the client wanted.”
If the results of an analysis don’t support the client’s message, said See, then “don’t impair or corrupt the analysis.” Forcing the results to show something they don’t will send the analytics team into a “death spiral,” in which they have to keep bending the rules. He also warned analysts not to tell clients they are doing their job wrong if results highlight problems.
- “Data does not equal analytics, and analytics don’t equal insights.”
Thorpe said data are elements to be analyzed, and analytics are “smoothed summaries” of data. Insights are unknown, surprising, and sustainable.
- “Is your analysis driving compliance or insights? Not knowing the difference is disaster.”
Thorpe said “compliance” is about checking a box that something was done. It’s avoiding risk. Insights can help the company succeed.
- “Great work does not overcome being late.”
See explained that businesses have deadlines. If an analytics project extends beyond a deadline, it’s crucial to be able to provide the business with as much information as possible. The deadline and the required decisions can’t wait, even if the final analysis is great.
- “The business will make decisions with or without you.”
Extending on what See said about being late, Thorpe noted that the analytics team is not more important than the final business decision.
- “What makes analytical jobs sexy is not the elegance of the math; it is helping the business win.”
Thorpe, citing a Google quote about the “sexy” aspect, emphasized that analysts can’t lose sight of what success means to the business.
- “Good analytics can’t overcome stupid decisions (or processes).”
See emphasized that analytics is only one element that goes into a decision, and that the analytics results aren’t the decision. “There’s always a majority of other factors at play.”
- “All experiments and analyses run into unforeseen problems; it is part of the game.”
See emphasized that things will go wrong in a project; learn from those problems. If you expert to run into problems you won’t be disappointed.
How to get buy-in for your program
Connolly also offers some great tips on how to get senior leadership to buy into your data concepts, which he says is cited as a challenge more often than any other he’s encountered recently. I’m guessing that a few of you have had similar issues. Here’s his advice:
- It helps to have a data advocate in the business unit, or backing from the CEO.
- Good data visualization is key. Without it most people have no idea what all those numbers mean.
- Show the good things data can do, because there are a lot of negative stories out there about using data to spy on people. Don’t just deliver a report, tell a story and be excited about what you’ve learned.
- Take a skeptic to lunch. Rather than just throw an email out there saying “This data is available, let us know if you have any questions,” sit down and break bread with teams that might use it. People are more likely to buy into your ideas if you are sharing food. Explain how other employees or other companies are using the data. There is nothing like a little competition to get people’s attention.
- Treat data like a frog. As Connolly explains: “Think back to your school days. What was more interesting, the teacher droning on about sentence structure, or show-and-tell day when some kid brought in their pet frog?” Maybe it’s time to bring show-and-tell into the world of analytics.
Thanks to vuralyavas at Pixabay for the image.