In last month’s Keyword, we explained why companies should focus on the quality rather than the quantity of data in business decision-making. To follow up on that thought, I wanted to bring up my view of the single largest obstacle preventing many digitally-driven companies from fully integrating analytics into their decision-making process.
On a granular level, the amount of data we are currently able to collect is unprecedented. Digital disruption has brought forth new methods of collecting, tracking and automating data, which are affecting businesses in all verticals. And although most companies' core offering isn’t strictly driven by data, their marketing efforts are – or dare I say, should be.
In all fairness, most modern businesses do use data to support their day-to-day decision-making. The era of blindly relying on the creative director’s gut feeling is long gone, and even CEOs have to leverage data to back up their strategic moves. Today, most companies use analytics to figure out which customer segments to target and which channels to use. Many businesses have supplemented macro conversion tracking with micro conversions, speading up the operational feedback process and ensuring that they understand user journeys on a level that matches the granularity of the data. Then how is it possible that so many marketing campaigns, products and services still fail?
The problem exists on two levels. To some extent, digital products and services are governed by the same uncertainties that shape the media industry. I’m referring to two properties in particular, namely “nobody knows” and “infinite variety”. While the nobody knows property is rather self-explanatory, the infinite variety condition refers to the unlimited possibilities of small changes that can be made to the design of a product or service. (If you wish to know more, I urge you to read David Hesmondhalgh’s book, The Cultural Industries.)
While new streams of data help decision-makers navigate and partly circumvent these properties, most industries and companies are still crippled by them. Furthermore, the internal culture of many businesses is built around a higher purpose or a cause. Although these deeply rooted belief systems drive employee motivation, they can also cloud judgment. When widely accepted “truths” are contradicted with new data, people tend to stick with existing, simple heuristics.
A great example of this comes from my own research on the use of analytics in the music industry. In the study, I looked into the data streams that a record label analyzes to predict the success – or failure – of newly signed artists. When asked point blank, a record label executive denied that he had ever received a bad outcome of such analyses. However, the known probability of breaking even on the initial investment of signing a new artist is only 10 percent, which means that a large majority of these analyses should have come back with negative results.
This glaring inconsistency implies that a decision to sign an artist was made long before the record label analyzed a grain of data. In other words, the company was only using analytics to back up and tweak their pre-made decisions. Although it may seem hard to believe, there's a logical explanation: it’s difficult to abandon an investment with such high sunk costs. Instead, the approach allowed data-savvy employees to support their opinions and frame results based on personal and organizational beliefs.
Reducing the bias
Much like the record label, many companies still base their decisions on the senior managements' experience and expertise, rather than predefined KPIs. The only viable long-term solution for reducing this bias is to standardize the analytics workflow, and implement clear guidelines for collecting and using data. Luckily, technology is helping businesses to visualize information, thus leveling the playing field for those who are not familiar with advanced analytics. Finding the relevant KPIs and investing in organizational learning to understand the impact of analytics are crucial for adopting a truly data-driven decision-making process. Such a shift in culture will need to start from the top of the organization, which can be tricky since there is often a disconnect between the key decision-makers and the employees who are in direct contact with data.
Data, analytics and insights are tools that can be deployed to reach business goals, and as most tools they are only as strong as the party that wields them. Data-driven decision-making is not about killing creativity or vision. Quite the contrary, it’s about supporting them. It helps companies to recognize when to kill their darlings, and when to pivot. The concept of “growth hacking” in which data analysis is used to build a product from scratch explains why companies like Instagram and Airbnb became successful – they pivoted based on data, even if that meant abandoning a number of early ideas and business models. At Luxus, we understand that data is not just data, and that data-driven decision-making takes more than just tools. It requires a deep understanding of the organization, its stakeholders, culture, products and industry.