As we have moved towards an economy that’s powered by data, the clamor for it in more and more areas has grown. Where it was previously seen only in boardrooms and amongst data teams, it is now the lifeblood of most successful companies.
This has come about due to its availability increasing across almost every department. Marketing now have access to huge swathes of customer data, finance can use it for more accurate forecasts, and even operations can use it to predict stock levels. It has put data into the hands of everybody, which has spread the benefits across the entire organization, creating a more informed and therefore better performing workforce.
It is not only in larger organizations where this change has been seen though, we are also seeing smaller businesses and even startups jumping on the data bandwagon. This is because democratization has also seen a growing number of companies start to offer SAAS data analytics services, some even offer data platforms on an open-source basis. It has meant that one of the key advantages that large companies had a decade ago is available to everybody, even those being founded to disrupt these incumbents.
Startups may not have the same kind of data sets that established companies do, but the ability to quickly scale this in this digital age is the very reason that the concept of big data exists in the first place. After all, who would have thought that within 22 years of their founding a company founded in a Californian garage would hold more information on US citizens than their own government? To some extent the need to build a database from scratch can be an advantage for startups as they know that the data they are collecting is up-to-date and relevant for their customers.
However, these two previous elements together cause a bit of an issue for companies, because as the use of data has increased, so has the desire to collect and use it. Therefore, as more departments and businesses begin to see the benefits of using data for their specific needs, the pressure put on companies to produce more analysis for every department increases. At the same time there is a reason that data scientists have a median annual salary of $116,840 - demand is currently far outstripping supply. So people want more data and data science departments want more qualified people to help satisfy this, yet struggle to find them.
This issue can often lead to people trying to ‘hack’ and utilize some the many other data software options available to them in order to try and emulate the same results they had from their data science team. So as these software platforms have become more accessible to startups, they have also become more accessible to everybody who wants to use them. In a large company, this can lead to losing the central point of truth that is essential to any data science program, or it can lead to the wrong analysis as those undertaking it are inexperienced and untrained.
Brent Dykes wrote about this phenomenon in Forbes, comparing it to his children playing in the shallows of a swimming pool, seemingly fine, but nearly drowning when they became confident and tried to use the deep end. The same can happen with data, the damage that can be done through untrained people using it in the wrong way can be significant.
It makes it imperative to manage your democratization, not simply let your company run free with the data they have. CDOs need to make sure they are aware of how the data is being used, who is using it, and put in place rules about how to use it. This can often be difficult to balance, but to quote a common saying, it is important to make sure you are giving employees rope, but not quite enough for them to hang themselves with.