Summary: Business instinct | When sums add up | Data-driven decision patching
This is a story about companies who like aggregations a bit too much. Data-driven decision making seems to be the new holy grail in management, but can the numbers always be trusted? What is key in data-savvy businesses: the people, the right technology, or – spoiler alert – is it something more fundamental? These questions become particularly urgent in the new economy as failing to embrace data can be a major growth impediment or worse, a dead sentence to the business.
Recently, I’ve had lunch with a friend whose job is to manage sales people. I enjoy our work-related talks: they nudge me to look at problems from the management perspective, so different to my usual consultant goggles.
My friend told me how a fellow manager of his prefers to trust her gut rather than the numbers.
“Oh, dear,” I heard myself saying. I could feel the Data Science fairy getting another bald patch.
But his argument was far from insane. He said, in a global company like his, reports you see can be more misleading than helpful. Managers often work with totals: total sale, budget, or hires; numbers that on their own are deprived of context. As aggregations are made levels above a single sale, the conditions that permitted these little occurrences to happen are discarded. Besides the typical descriptors like region, product, or salesman, there are few known details. The information is too scarce to fully elaborate on the underlying strategy, what worked and what haven’t. Management is deemed to operate in this highly generalised vision of business.
In that manager’s view, totals are too out of context to be trusted. With that level of generalisation it’s easy to lie through numbers, or to cover up a lie. For her, following your instinct in decision making is just as valid a strategy as any. Sadly, and with a feeling I’m betraying my work principles, I agreed. She wasn’t fundamentally mistaken; the contrary, she saw right through the wiggly foundation of a system that only appeared legit.
My friend and I agreed that this strategy was unsustainable. Zara came up in the conversation: “These guys really know how to use their data.”
Zara is a massively successful global company. Zara redesigned their operations to respond faster to market: trend, production, and delivery-wise. One of the key elements of making this strategy work was stepping away from centrally-planned collections. Zara realised that their Spanish branch caters to other customer needs than Zara China. Market differences were the force behind restructuring the organisation. As a result, the decision power has been distributed across local entities. Regional management was given only its own market to worry about, hence its response to changes has become faster, more relevant, and more agile than any central body could produce. Zara appreciates market diversity; a single, global collection won’t fit all.
This is when it clicked for me: the real question wasn’t about what analytics is or isn’t able to do, but about how it’s being applied by business. In my friend’s professional life the managers mistrusted Business Intelligence, while companies he admired embraced it wholeheartedly. Analytics, in theory a key enabler to informed decision making, for some organisations plainly backfired. These companies saw the main trends, but were oblivious to the multifaceted nature of the collected data.
Was the technology to blame? We all know the opposite: today, classic analytics combined with the sophistication of the visualisation tools and the data processing engines give us unprecedented access to data. Business Intelligence thrives in multidimensional information analysis: viewing data from different angles is its building principle. One angle could be reviewing sales on a regional level and comparing it to a moment in time: how much did we sell last month in France? Is that an increase from the same month last year? A more sophisticated angle looks at data distribution in regions within an area to see discrepancies in the buying behaviour across geographies. Another analysis studies the buyers demographics by comparing it to what is known about the general population and about the competitor’s base (as much as is legally allowed). Formally called “slicing and dicing”, it has been a feature of the Business Intelligence systems ever since there were Business Intelligence systems.
Potentially a game-changer, analytics are universally misused. BI reports might be produced with meticulous care, but are understood only by few. Managers might appreciate the value of analytics, but are untrained to work with data and unable to ask the right questions to challenge somebody else’s product. In result, the technology is misinterpreted, misapplied, or at best used to back up somebody’s gut feeling. Analytics, however brilliant, is no magic bullet – and unless the business culture changes, it’s no bullet at all.
The problem is systemic. The reality in many businesses is that the decision power is highly concentrated. These few decision makers operate on a flattened version of their business: analytics lends them the correct data, but its opaque in detailing its constituents. The data is fed back to lower management once the strategy has been decided on the top level. The top-down communication lets people on the view of the management, but gives no authority to conclude their own view. They therefore make decisions based on somebody else’s perspective, or as the aforementioned manager, they reject the irrelevant data and go with their intuition. This business culture fails at empowering the merited decision makers to act on available data. The system is designed to kill agility and forego context.
Data-driven decision making is the new catch-phrase among the management, but how is the data supposed to drive change if the system isn’t designed to allow change?
Data-driven decision patching The world is full of examples of companies who love aggregations so much, they stop paying attention to what happens at the point of sale. Many of these companies are global. Economies of scale they benefit from, technology barriers, and likely little competition, have made them tone-deaf to their customers.
So-called “customer care” is the number one example of business ignorance. Most service providers (mobile networks being the flagship example) allow the customer to only deal with the first line support. There are armies of people trained to tell you that “this-is-all-they-can-do”, or that “this-is-the-company’s-policy”. Rarely do they provide more help than a Google search, but you get an impression you’ve been listened to. If they cannot solve your problem or upgrade your offer, then you take your resentment home: the case is not even logged. The management does not know about you. The company never changes. Yet, their aggregated sale numbers still show a profit.
Big Data has seemingly come to change the status quo. Customer lock, i.e. unlikeliness to switch a service provider has been identified as a massive hurdle to the business growth. Getting a new customer is said to be expensive because it requires snatching the person off the competition. Keeping the customer is relatively cheap because he/she is already with us but – coming back to point one – it’s very costly to lose one. Many companies leverage Big Data to identify potential churners and stop the loss. People’s behaviours are analysed: how are they using the service, how are they paying for it, are their friends using the service? This is clearly a step up from reporting, but again it only continues the same stale way of making business. There is no effort to change the company. Identification of an unhappy client and a resolution that could have happened at the customer care level is delayed and dealt with late, if ever. This is data-driven decision patching, not making.
Business culture needs to go through a fundamental change to become data-driven. It’s not the new tools, more data, or a PhD-holder staff that will make the change happen. A system, a program, or a robot, designed to be most rational, capable, and failure-proof is put to action only if the environment it operates in allows it. Analytics have been available for years, but they continue being misused because the business models they aid are inherently incompatible. Centrally manged organisations by design impede change; its time that decision making becomes more democratic and distributed, as is data.
If you liked this post, make sure to check my last article in which I look for the elusive Data Scientist, and follow me on Twitter for the updates and some random tweets!Follow @EveTheAnalyst