
In the old days, about 3 years ago, when a manager wanted information, he/she’d sent a request to the IT department. In a few days or weeks, the manager would have the information needed for decision-making. Some businesses still do data analytics this way. But, according to a recent survey by Intel,
… speed is key to delivering successful big data analytics. Speed is not just a challenge with how quickly data is being generated, but it is also a matter of how quickly it can be accessed, analyzed and acted upon to have a clear impact on real-time (or near-real-time) decision-making and business responsiveness.
What used to take weeks, now must happen at the speed of light — real-time analytics.
For example, WellPoint, a health benefit company, found that it could take up to 72 hours to respond to a doctor’s urgent request for coverage of a treatment. In an effort to slash that time, the company implemented a system that quickly correlates clinical research, patient data, and clinical practice guidelines to expedite the pre-approval process on doctors’ requests. The new system generates hypotheses and uses evidence-based learning to score recommendations on a confidence scale so nurses are given the best options for each patient in a matter of seconds, not hours or days. By shortening the approval time for treatments, WellPoint is reducing costs, improving service, and, most importantly, singling out the most appropriate treatment for each patient. – by Information Week
Data analytics at the speed of light
While IT professionals tend to point to legacy systems as the big bottlenecks in providing real-time analytics, the impact of technology is dwarfed by cultural impediments. You can buy new and better stuff, making a cultural shift is like moving the Great Wall.
Here are the true impediments to doing data analytics at the speed of light:
- bloated bureaucracies
- faith in the technology behind the “black box”
- decentralization of the information function
- investments in highly skilled analysts and managers
- internal politics
Let’s take a look at these 1 at a time.
1. Bloated bureaucracies
Requiring managers to get buy-in from multiple layers of managers for decisions slows the process to a crawl. It doesn’t matter how fast you acquire data and process it into usable insights, if the decision approval process takes days.
If you don’t trust managers and analysts, then you’ve hired the wrong people. Get rid of them and hire employees you can trust. Then push authority down so these managers can make data-driven decisions. Provide oversight, not a roadblock.
I find that businesses too often view hiring skilled employees as a cost rather than an investment. We’ll talk about this later.
2. Faith in black boxes
Often, managers lack the understanding to place faith in the results coming from a “black box”. So, instead of making data-driven decisions, they make decisions based on their own instincts or place their faith in how it’s always been done.
A survey by IBM found 63% of businesses across 70 countries who invested to improve their data analytics saw positive ROI within 1 year and 69% of companies focused on improving the speed of data insights got positive returns. Hence, data-driven decision-making outperforms every other type of decision-making for most companies. Companies like Amazon and Google who are the poster children for data analytics, nearly every decision is data-based and these firms see much higher returns for their investment.
Two things are increasingly likely over the next few years:
- The ROI difference between firms using data analytics and those not will increase significantly
- Data analytics complexity will make it harder for managers to understand the mechanics of what’s happening within the decision-support system. As machine learning and other algorithm based analytics tools arise, intuition won’t aid understanding how they work
In graduate school, 2 camps occupied the analytics space — 1 favored hand calculations to support the findings offered by increasingly sophisticated software programs, the second favored increased understanding and interpretation of the RESULTS from these programs. Obviously, I fell into the second camp and still do. You don’t have to be able to calculate a correlation coefficient or Beta in order to use the information they provide for decision-making. You trust the folks who programmed and tested the software to have done it right. The distinction between these 2 camps is more salient now than it was then. Now, machines can learn, as we saw in the WellPoint example above. They make routine decisions faster and more accurately than humans, which provides serious returns.
If you or your manager are skeptical, trying running experiments to test the results of data-driven decisions against those coming from other decision-making schema. Try running a division, product line, or campaign then comparing the results rather than just toss out the data-driven recommendations because you don’t understand where they came from.
3. Decentralize the information function
The distinction between and analyst and manager has largely disappeared. Some, like IBM, recommend replacing it with translators, such as we see in the exhibit below.
I’m not a big fan of this as it increases the layers of bureaucracy between data and action.
I’m also not a big believer in a separate data analytics department.
If I were constructing data analytics within a company (which I’ve done many times for clients), I would use cross-functional teams to manage the company based on project, products, geography, or whatever makes sense for the particular organization. A data analyst is an integral part of the team, providing insights and answering queries. Each data analyst works with other analysts within the firm who are working with other cross-functional teams to build a meta-analysis that ensures insights translate across the entire organization.
Analysts can’t be the guys from IT, as in the past. Instead, they must be business professionals who understand business processes. Improved interfaces and an emphasis on BI (business intelligence) within business schools means more folks have the necessary training to be business analysts (not FINANCIAL analysts).
4. Invest in highly-trained analysts
But, these guys (generic, not gender specific, LOL) don’t come cheap. Think of their salary (and continual training) as part of doing business because your competitors sure do. Go cheap and you’ll leave money on the table (through lower income) and consistently fall behind vis-a-vis your competition.
Let’s look at an example.
Assume you’re a Fortune 500 earning pre-tax revenue of $100 million. According to the IBM study, on average, you can expect a 1% increase by using data analytics effectively. That’s an extra $1 million every year. That’s a LOT of money you now have available to pay analysts. Maybe you’re an SME with an income more like $15 million; 1% of that is $150,000, still enough to hire a pretty good analyst full time.
But, lets not stop there. Let’s look at the opportunity costs of NOT hiring an analyst, resulting in higher operating costs, poor quality and poor customer service. Maybe that opportunity cost is $100,000, not unreasonable for an SME and much larger for our Fortune 500 company. Now, the cost of NOT hiring an analyst is $250,000, which could get you a pretty top-notched analyst.
5. Internal politics
This isn’t anything new.
Internal politics has always been a roadblock to effective management. Managers might KNOW what to do, based on data analytics, but lack the WILL to do it. Maybe the decision will shift resources from a more powerful manager. Maybe the decision has a substantial cost, despite a significant payoff in the end.
Any number of political issues stand between data analytics and good decisions.
To survive in the data era, firms must strive to eliminate such politics from the decision-making process and rely on data-driven insights to generate appropriate decisions.
Need help?

We welcome the opportunity to show you how we can make your marketing SIZZLE with our data-driven, results-oriented marketing strategies. Sign up for our FREE newsletter, get the 1st chapter of our book on digital marketing analytics – FREE, or contact us for more information on hiring us.
Hausman and Associates, the publisher of MKT Maven, is a full service marketing agency operating at the intersection of marketing and digital media.