The following is a guest post by Joy E. Taylor, co-founder and CEO of TayganPoint Consulting Group
What matters most with Big Data? Turning it into something companies can really use.
Big Data has always been a kind of siren song to executives, as if it held the answers to all their problems. “If only we could get our hands on more data,” they insisted, “we’ll make better decisions sooner.” Trouble is, we’ve had access to more data in the past 15 years than all previous years combined, but we’re not making decisions that much faster.
One reason may be that pulling bits of useful information from Big Data has lately become like finding a needle in a haystack – and as the haystack grows, that task gets harder and harder. Nassim Taleb, author of The Black Swan: The Impact of the Highly Improbable, looks at this in his latest book, Antifragile. Taleb’s theory is that as more information is available, people will unwittingly find fake statistical relationships; they are, in a sense, fooled by randomness. To make matters worse, those false correlations could have huge implications for companies basing their decisions on cherry-picked or erroneously valued information.
This may be one of the largest potential pitfalls of Big Data. When you have an ocean of information at your disposal, it’s tempting to keep fishing until you find the facts that support your bias or agenda. That certainly does nothing to support agile decision making.
Yet it’s not like CEOs are giving up on Big Data and analytics as keys to improving business performance. McKinsey & Co. found that more than a third of C-level executives believe Big Data and analytics, social media and other digital business tools will increase operating income by 10 percent over the next three years.
Fix the Way You Access Information
It’ll take more than believing in Big Data. In my consulting work, I’ve found that while our access to data has changed, the structure of our decision making hasn’t. In most organizations, when executives want data – performance metrics, financial results, etc. – they usually go to finance, which pulls up data, works it up into some kind of visualization like PowerPoint slides. Then that information is reviewed and cleansed for three to five days. So by the time it reaches execs, it’s days, even weeks old.
As a business process, this is patently insane. It keeps people from making timely decisions because it delays their access to critical information. And you can’t realistically just “will” yourself to overcome it. If senior execs at a certain level are motivated, they have to work within the constraints of their organizations using decision-making structures established not just before Big Data, the cloud and mobile, but before the digital age itself.
To fix this, we have to put in place the kind of structures that give people the confidence that the information they’re accessing is accurate and worthwhile. Say you’re not feeling well. These days you can just enter a few symptoms into your iPhone to get a sense of what you might have. If Google gives you several similar responses to those symptoms, then you’ll have fairly high confidence that you’ve identified the problem. But if you get 19 responses with all kinds of different possibilities, your confidence plummets. In today’s real-time business environment, we can’t afford to get 19 wildly different responses to our queries. We need the one right response, every time. We need to find the right needle, to separate the signal from the noise.
You Gotta Have Context
This is where context comes in. My longtime friend and collaborator Christian Gheorghe, who is founder and CEO of Tidemark, has thought deeply and written extensively on this topic. Christian argues – correctly, I believe – that data without context is confusing and often useless or even misleading. And in the ever-increasing data haystack, context matters more than ever. Without context, it’s impossible to assess the value of data, or to appreciate its potential implications.
Looking at this from an enterprise perspective, I can see situations where managers might look at a handful of data points and draw inaccurate conclusions from them because they either can’t access the context around that data, or they simply root around in all that Big Data until they find the answer that confirms their bias.
For instance, let’s say an electronics retailer operates a service and repair arm. Management has decided its Kansas City service facility is overstaffed. They’ve concluded this because Kansas City closes 20 percent fewer job tickets during the typical workday, compared to other facilities with a similar sized staff. They’re assuming the workload and the staffing levels don’t match up. That’s a fair assessment until you dig into the context and find out that Kansas City is the only facility that accepts walk-in orders. And digging further – perhaps by looking at email conversations, memos or other pertinent contextual information – you find that the average service employee loses 2.25 hours of repair time every day because they’re processing walk-in job tickets and sitting on phone calls from frustrated customers. Turns out Kansas City isn’t overstaffed – if anything it’s understaffed, or at least its staffing levels are in dire need of load balancing.
Use a Standard Unit of Measure
Enterprises have a big advantage over consumers in that we control where we get our information. And with the right technology or tools, we can control how we’re digesting and using that info. But incredibly, much of our data is hard to access, hard to use, and hard to analyze because different people and functions view it differently. They’re not using a standard unit of measure to talk about metrics, and that slows down decision-making.
But I think it’s fairly simple to achieve this. You start by identifying how the organization talks about its performance to the street, to investors, to board members. Then make sure we are pulling data using the same terms and parameters that we use to communicate to whoever the company answers to. This ensures everyone – finance managers, line of business managers, C-level execs – must share the same operational definition.
Take FTE (Full-Time Equivalent). FTE can mean different things to different people. But when finance gives metrics on FTEs, you can bet that metric is specifically calculated the same way, every time. Does everyone know this? Does HR? Do LOB managers? I’d argue they don’t. And you can immediately see the problems arising from that. Imposing a standard unit of measure on all reported metrics is essential, otherwise when two different people or departments are looking at FTE, they’re not looking at the same information.
Fix Your Technology, Too
It’s not just an organizational problem. Modern data gathering and analysis relies heavily on technology. Just ask yourself where data relating to COGS (Cost of Goods Sold) resides. The answer is in many places – sales, manufacturing, purchasing, freight and shipping, even facilities. But data that feeds the COGS metric also feeds other metrics. In a properly structured environment, data points that feed all standard performance metrics should automatically flow into all relevant metrics.
No hardware or software can solve the problem of an outdated decision-making structure, but once a strong structure is in place, the right technology puts that structure into overdrive. My 10-year-old daughter sells Girl Scout cookies. I can walk into my garage and see her remaining inventory: she has three extra cases of shortbread cookies and two of Thin Mints. So in 10 seconds I know she needs to push shortbreads and Thin Mints on her next sales outing.
Performance of all kinds improves when you arm people with the right information at the right time, and then make it possible for them to act on it. Whether it’s enabling sales representatives to know what they need to focus on that day, or helping HR to reduce average hiring times by understanding which positions are harder to fill and analyzing how to fill them faster — data makes this possible, but enterprises also need the right structure and the right tools to bring that vision home.
It’s Time for Companies to Catch Up
Those wistful, halcyon days of waiting around for more data are long gone. Every business now has access to virtually all the data it needs. New groundbreaking solutions, like cloud-based enterprise performance management applications, make it possible for companies to rein in data, sift signals out of the noise, and focus on only relevant information.
I’m seeing this happen among the companies I work with. CFOs and COOs are using these new platforms to create new company cultures that pushes decision making down from the corner office and out to the edges of the company where the day-to-day work gets done and the data itself is generated. These solutions once were the sole dominion of a few highly-trained finance wonks. Now we have companies like making them available to managers throughout the company using virtually any mobile device. It’s a matter of redefining who the decision maker is, of cutting out layers from the decision hierarchy so people have the insight to do the right thing at the moment they should be doing it.
The technology is ahead of most companies on this, so businesses had better catch up. Fifteen years ago businesses asked for this environment, and now it’s here. It’s time to really do something with it.
Joy E. Taylor is co-founder and CEO of TayganPoint Consulting Group, where she uses her formidable Six Sigma Master Black Belt skills to help build high-performance organizations and shepherd major projects and initiatives.