By Matthew Robinson, Milestone Partners
As part of ACG Philadelphia’s Industry Vertical Roundtables, four data experts engaged in a fascinating, insightful and entertaining discussion on the importance, relevance, and changing dynamics of data analytics today. Colin Hare, Managing Director, Strategy and M&A Services, KPMG moderated the session.
No more playing fast and loose
The market is seeing a ton of behavior change and people are shifting how they perform data analytics; this was the general consensus of our featured speakers – partially in response to GDPR (effective May 25th) and partially in response to recent headline data leaks (Facebook, Target, Ashley Madison). Kicking off the discussion, Mr. Stein touched on the evolution of how company data is viewed: it was previously seen as an unambiguous strength to have a plethora of data – now, a selling point for his company, Stitch, is that they are able to automatically set data retention policies for other companies (one requirement for companies to be in compliance with GDPR is to delete data promptly in response to user requests). Companies don’t know what to do with all of their data, and they are afraid that it could end up exposing them to liability. The roundtable discussed two examples of market consequences that may have been noticed by the common eye: 1) an internet search engine by the name of DuckDuckGo is gaining market share by guaranteeing searchers’ privacy (and ultimately reducing targeted ads) and 2) many eCommerce sites now re-direct customers to another page when executing a purchase to outsource the liability of receiving personal and credit card data. As companies scramble to protect themselves from a data mishap, they are actively looking to shrink the amount of data they possess that people might want to hack.
Two clicks to insight
A reoccurring theme in the roundtable, identified by Mr. Leininger, was the idea of “two clicks to insight”. IntegriChain embraces this motto to make sure its data scientists focus on achieving real insights and not losing people along the way. If a model is going to take more than two clicks to get to an actual insight, then don’t build it. Nowadays, there are data scientists who can write code in 10 different languages and run thousands of tests simultaneously – in a matter of seconds. While impressive, data scientists must recognize that the business importance lies in the insights, not the code that got them there. Data analytics reports can be of endless length, but the reality is that the majority of the information doesn’t matter. Senior executives are not going to look at a report if it takes them more than two clicks to get to an insight.
Focus on the output
Mr. Leininger emphasized a key theme: “The only way data can be a strength is if something actionable comes out of it.”
Companies are struggling to figure out what to do with the extensive amount of data that they have – it doesn’t need to be overcomplicated. To the people who say they have data and don’t know what to do with it: the panelists would advise you to find the 2 or 3 bottlenecks in your business and reverse engineer an insight that is actionable. Don’t let the massive amounts of data get you underwater. Think to the problems in your business that you are trying to solve, and then think how your data can help you solve that problem. It is important to remain focused on finding the one or two most important things in your business that you can impact with data analysis.
Mr. Angert cited an interesting example from the early days of Facebook. With a lot of smart data engineers and a ton of data (maybe too much…), Facebook was able to develop their well-known “7 friends in 10 days” rule. Basically, Facebook figured out that they had an incredibly high retention rate among users who reached 7 friends within 10 days of making an account with Facebook. When the company realized where people were most commonly falling out, they shifted their focus to engineering different processes that would result in a higher percentage of new users reaching the targeted milestone. This is a prime example of a company identifying an actionable problem and applying data science to fix it.
The speakers agreed that often times the less experienced data analysts present an overwhelming number of charts and graphs, and the more experienced analysts show two or three conclusions designed to drive strategy, all supported by the data. Mr. Leininger added, “Don’t just show your board a bunch of numbers. You need about three.”
Should my company be performing data science?
If CEO’s aren’t driving data science at their company, they are not doing their job. Typically, somewhere between 50-100 employees is when companies get enough data/scale/questions to bring in a full-time data scientist. Before that, the role is simply assumed by an employee who knows their way around excel. A common catalyst to creating a data scientist role is some sort of reoccurring question that the company can’t answer – most commonly driven at the Board level. When the CEO and CFO get tired of having to say, “I don’t know,” they proceed in hiring a junior data analyst.
An alternative (and cheaper) option that some companies rely on is an outsourced data analytics model. Mr. Angert noted that he is not a believer in this alternative – in his opinion, you need someone close to the business to be asking the right questions in order to gain insights from the data that are valuable to the company. Shipping off your data to a consulting firm and receiving differentiated insights that could give your company a competitive advantage is highly unlikely.
“In God we trust; all others bring data”
If company decisions are left to the gut, people will inevitably defer to the management org chart. Left with a fascinating W. Edwards Deming quote, “In God we trust; all others bring data,” it was clear to the entire room that the days of making decisions with your gut are near over. At a time when industries and companies are more competitive than ever before, the importance of making data-driven decisions is emerging. If you want your idea or position to be supported when presenting to a room full of intelligent people, it is essential that you bring data to back it up!
Thank you again to our sponsors