Just about every business uses some form of data in making decisions, but not every business has a data-driven culture.
So what’s the difference? The main distinction is that a data-driven culture flows from an executive suite that has committed to a data-based approach. Strategic decisions rarely (if ever) are made without incorporating some sort of data analysis: some examples are analytics, the results of customer surveys and conversations, experiments, purchase histories, and even location data.
For example, a data-driving marketing company will have an entire suite of tools logging as much useful information as possible for later analysis. This will probably include analytics tools that show how customers find their clients (the search terms they use and other sites they click through from), customer relationship management tools that extract useful data from regular customer interactions, and “heat maps” that show what aspects of a site or an ad the customer is most interested in. The data generated by analyzing all these inputs directly informs business decisions.
The data-based approach is even used in fields that aren’t sales-oriented, such as journalism. Data-driven reporting tackles complex situations with lots of variables by attempting to define and measure all of them using many types of data sets: government reports and studies, financial disclosures and performance data, interactions on social media, and so on. Some academic fields, such as sociology, employ a similar approach.
The movie “Moneyball” is a good analogy for the greater shift to the data-based approach in business. The central conflict of the movie is the introduction of a data analytics strategy to a culture that had previously operated by an approach to decisions that relied on the intuition and personal anecdotal observations of managers.
Prior to the introduction of “big data” this is how decisions tended to be handled across all types of business; a manager or executive was often appointed based on prior experience handling a similar situation. However, that prior experience might have been years out of date and no longer suited to the current circumstances.
All of this may sound somewhat simple, since many organizations have incorporated data analysis to at least some degree. So why are so many not getting it right?
Data-Based Cultures: Easy to Sell, Hard to Make Stick
It’s not that hard to articulate the benefits of a data-based approach, since so many organizations have onboarded some elements of it already. The challenge is in getting it to function across the organization.
It can be tough to get employees to buy in to the new mindset, particularly if the organization has a long history of managers “going with their gut.” The greatest influence comes from the top. If senior management is invested in making most (if not all) decisions based on data analysis, employees will need to do so as well to keep pace.
Other common issues boil down to communication dysfunction. Data analysts and scientists are not integrated with executives or the rest of the business. Collected data is not easy for employees to access, or they have not been trained in how to make use of it. “Cliques” that prefer different techniques or metrics or programming languages form and stop communicating effectively with each other.
The Fundamentals of a Data-Based Approach
Encouraging buy-in is helped tremendously by having a clean, functioning model that demonstrably works well.
The specifics of this will vary somewhat by organization and mission, but there are some universal expectation and requirements for data-driven decisions:
- Identifying data sources and coordinating them with present and future company uses
- Cleaning raw data (removing items that are not useful or not complete)
- Building analysis models used to glean important insights from the data
- Creating processes to effectively communicate these insights throughout the company
Benefits of a Data-Based Approach
If a data driven organization is doing things the right way, it should see both a major improvement to the efficiency of existing processes and sparks of innovation that take things to the next level.
In the broader sense, success can be predicted by using data rather than relying on the intuitive skills of key decision-makers, making for a more stable and consistently prosperous company.
It’s also a general morale booster. Rank-and-file employees no longer feel they are working at the capricious whims of the boss from Dilbert, but a collaborative process that anyone can verify for themselves is successful.
Implementing a Data-Based Approach
While it sounds like a no-brainer on paper, there can be serious obstacles to implementing a data-driven culture and it is not something organizations can be expected to pivot to overnight.
There are some fairly predictable challenges that can be anticipated:
- The replacement of legacy systems, which members of the company may be deeply attached to
- Members of management that feel threatened by an approach that shifts away from autonomous decisions
- Existing issues of inter-department communication dysfunction that will hamper access to data
- Implementing any necessary new training programs
While none of these issues are insurmountable, they can take considerable time to overcome. Organizations generally get the best results by starting small and implementing predictive models in one area of the business operations. Establishing a record of success in one area and some familiarity in the culture makes it easier to gradually spread a data driven approach through other areas.
Importantly, the organization’s analytics team can improve their decision-making process by ensuring that the data they use has the necessary breadth, depth and fidelity. Not all data is created equal, so investing in the exploratory efforts on the front end of the process will pay dividends on the back end.
For example, if you are considering using third-party data to support marketing efforts of your new consumer facing product line, your data providers must have an expansive methodology to account for any potential biases in the information you are using, while also having enough depth to ensure you can extract the value you require. Clickstream data is one type of data that can be helpful due to its passive nature of objectively measuring user behavior.
According to Eli Goodman, CEO & Co-founder at Datos, “Clickstream data is essential for companies serious about a data-driven approach for decision making impacted by online human behavior. By leveraging anonymized, privacy-compliant datasets, companies can avoid violations of data privacy regulations while still getting access to the desktop and mobile browsing behavior for millions of opt-in users across the globe.”
Are There Any “Cons” to a Data-Based Approach?
There can be temporary setbacks (aside from the complexity and initial expense), particularly if the approach is entirely new to the organization. For example, incorrect cleaning and interpreting of the data could create “insights” that aren’t actually very useful … or worse, ones that send the company in the wrong direction.
Another common issue is when a supposed “data-based approach” mutates into simply cherry-picking data to reinforce a preconceived conclusion rather than making informed decisions. This is generally a cultural problem, one that often stems from leadership refusing to truly buy into the approach or co-opting the process as a way to showcase personal (rather than organizational) success.
The answer to these “cons” is a commitment to collecting and deploying accurate data that everyone in the organization can have confidence in, ensuring that training is in place to ensure that everyone understands how to interpret data correctly, and ensuring that failures and unknowns can be openly discussed without individual team members feeling their career is at risk.
