First of all, here by data we mean processed data, possibly twice, so knowledge.
Beware: without this clarification we risk making decisions on raw or not refined enough data (descriptive statistics).
What differentiates relying on data versus being informed by data? The decision-making weight of processed data. Using a metaphor, in the former case the “Data” (processed) partner has the majority, in the latter case it does not. In other words, in the first case the (processed) data has the most weight in decision making, in the second case it does not. Insufficient statistical awareness may also lead to the second approach, in addition to some psycho-cognitive problems in the company.
The former have less need for managers, especially if statistical process learning is used and obviously this may bother them. The second ones “validate” statistical results more; managers or the owner have the final say.
Data-based (data-driven) companies, in my opinion, have no superiority over data-informed (data-informed) ones. So I don’t think there is a hierarchy but only contexts, and obviously in some of these using the second approach is limiting.
In what contexts should the second approach be preferred over the first? When the company does not have sufficient data quality, a word that seems abstract for now. Then you need to compensate with a flesh-and-blood figure, a manager or the owner, who, through his experience and/or intuition, completes the puzzle.
Certainly, in certain companies, there will exist departments that are data-driven and others that are data-informed. For example, a manufacturing company with an Internet of Things (IoT)-based production facility, which therefore produces rivers of data every second, may base production, but more so maintenance, on data. But production will depend on the business situation, and the techniques for having commerce. Here you have neither timeliness nor completeness of the data, components that define the quality of the data. So this type of company will have a hybrid decision-making approach, something some call (companies) inspired by data.
In some contexts you cannot have a purely data-driven organization because the European General Data Protection Regulation (GDPR) requires human beings to be present in the decision-making process. So “code is law” (in this case fully automated statistical analysis or models), a slogan coming from some blockchain circles in the year 2016, which led to community splits, does not apply.
Those who are youtubers or those who want to reach customers through content, such as Enterprise Statistics, can have the two approaches above on YouTube: the second approach has them look at video results to decide next video topics; the first looks at trends (e.g., with google trends) and indicates video topics based on that aggregated data (information). Plugging into trends (newsjacking) is thus part of a data-driven approach.
Beware that “data-driven” can become data gutted (data mislead). Also because of the damn cognitive limitations. For example, you want to decide which image to use for a post (not a campaign). The test statistic says image A brings more views in a statistically significant way. But companies need to find people who find their product/service useful and purchasable, not views. So the test gave a correct conclusion to a wrong or ill-posed question , unless the company verified that more views bring proportionately more sales. In other words, one must keep in mind the meaning of KPIs: key performance indicators. Views do not belong to the key indicators . As you see a single word can have very different consequences, just as in chemistry a single atom can make the difference between life and death (water and hydrogen peroxide joke).
Do you want to understand how to make certain areas of your business data-driven? Or how to have a hybrid approach? Let’s discuss this in a free call of about 45 minutes. Either way, knowing how to make data work for you becomes an important company asset, especially in SMEs.
For exercise, I’ll leave you with this situation: if a company creates and administers a survey to its customers, does it fall into a mode of having a data-driven or data-informed strategy?