It may sound hard to believe, but the term "data scientist" has been around for less than 10 years. Yet in that time, data science has become one of the hottest fields around, with every company worth their salt scrambling to keep up.
But the biggest question that companies aren't asking themselves is whether or not they actually need data scientists. This is not to say that what data scientists do isn't useful or doesn't create value for the company; rather, it's merely to point out that the industry has evolved, and there are now tools available that can perform some of the same functions with less hassle.
Because data scientists, like AI researchers, are so in demand, they often command extremely high salaries. According to Glassdoor, the average base salary for a data scientist at entry level or with minimal experience is over $100,000 a year. Realistically, most agencies can't afford to hire someone this expensive, especially given the fierce competition amongst established companies for qualified applicants. Instead of bankrupting themselves for the sake of a data science department of one, agencies are looking at other ways of getting the same results.
The demand for data scientists is rooted in the economy-wide embrace of big data. Without data scientists, the story goes, big data sits in a silo, slowly getting stale and losing all elements of usefulness. But what makes big data useful is not purely the fact that it's information that can be studied to gain valuable insights on consumers; it's also the role it plays in developing effective machine learning--and, by extension, deep learning--networks and algorithms.
Most brands and agencies don't have the amount of data available, nor the time, nor the resources, to build an effective machine-learning platform. This may sound harsh, but it's true; after all, how many brands or agencies have $100,000 to spend on a single employee, let alone the money for the platforms and software needed for that person to do their job properly?
Understandably, there are many out there who fear being left behind, and who are making significant investments in data science and analytics in order to prevent that from occurring. But the truth of the matter is that not every company needs a data scientist, especially given the proliferation of tools and platforms that can perform many of the same functions at a lower cost.
If you're a brand or an agency, it's nice to have a purpose-built AI model--but it's hardly a requirement for carrying out your day-to-day business. There are many existing platforms and machine-learning algorithms out there designed specifically for marketing use that make it less urgent for brands and agencies to own their own analytical infrastructure. In other words, not having a data scientist on hand doesn't mean that companies won't be able to take advantage of analytics platforms or AI.
"Big Data has leaped past what even teams of data scientists can hope to achieve with machine learning other than deep learning," says Jeremy Fain, the co-founder and CEO of Cognitiv, a neural network technology that helps companies make better decisions. "Deep learning resources are extremely scarce and will take a long time to build, for any company. Agencies and their clients will more quickly benefit from new technologies by partnering with companies that have already built the deep learning expertise."
Companies like Cognitiv should be considered as valuable partners for brands and agencies and their clients. While custom algorithms look like they are the future--AppNexus, the Trade Desk, and Beeswax all allow you to "Bring Your Own Algorithms"--the only ones able to take advantage of the gains custom algorithms bring are those with big data science teams and machine-learning platforms. Most of these companies are managed by services or platforms that have been building these teams for years. The next wave of platforms that offer these benefits to brands and agencies, instead of acting as a secret black box, will be the essential partners of tomorrow.
An article in the Harvard Business Review notes that companies now have widened their focus, "recognizing that success with AI and analytics requires not just data scientists but entire cross-functional, agile teams that include data engineers, data architects, data-visualization experts, and...translators." That's a long list of people to add to your organization in order to be successful with your homegrown AI and analytical initiatives, and it gives a sense of the scope of investment required.
The principle behind outsourcing your data science initiatives is, in many ways, analogous to the rationale that many large companies have for outsourcing their advertising campaigns. Not only is it cheaper to hire a company for a short period of time than it is to hire and maintain an entire creative department in-house, it's also beneficial to lean on the expertise of professionals with experience in the field and the know-how to produce successful campaigns.
Simply hiring a data scientist won't make you relevant, but sticking to what you do best, and relying on partners to help you with everything else, will.