Democratization of artificial intelligenceIt's rather clumsy that the profession of turning data in something useful is called 'data science'. Although a data scientist uses quite a lot of methods and tools that – up until a decade ago – were mainly used in academics, the data scientist is more of an engineer than a scientist. By combining programming, statistics and in-depth business knowledge, data science often solves problems and automates tasks that, until lately, were outside the realm of automation. These tasks usually involve some form of human judgement such as offering a quotation or classifying news articles. By training an algorithm with often huge amounts of data, data scientists can make computers mimic this human judgement – also known as artificial intelligence.
Since data science is rooted in statistical methods that are mostly associated with academics, it's not surprising that the early data scientists were mainly schooled in mathematics or hard sciences such as physics and astronomy and collected PhDs on the side. However, many applications are so self-evident that they are offered as off-the-shelve tools or as a service. With recommendation engines and labeling services, most of us don't even think about the AI technology that's under the hood. Furthermore, several cloud computing services and programming languages are optimized for processing gargantuan amounts of data, democratizing the methods used in data science.
The academics trapDespite this evolution, too often do recruiters and managers get hypnotized by degrees and academic track records. It's true, for research it's advised to have a firm background in math. Nonetheless, for solving business problems, it's often more effective to clobber your way through the corporate bureaucracy to get access to decision makers and the right data than knowing the probability mass function of a Cauchy distribution out of your head. Believe it or not, 90% of the time data scientists are collecting, cleaning and transforming data. Because in many cases, the low-hanging fruit is in good data and not in training a dozen of algorithms of which one is half a percentage point better than the rest.
The vagueness trapThis opens the door to a second kind of trap: the vagueness trap. Many consulting companies are attracted to the hype of the data economy. Remember when "digital transformation" was a vague notion of a website and social media? Well, history repeats itself: by quickly founding new spin-off companies or departments, agencies hope to attract companies looking for "data transformation". Some consultants get dropped at your company's reception, they will work on some opportunities that might be vaguely related to the data at your company or they'll do a workshop which results in a data transformation guide that basically says: "hire a data team".
In this context, it's often better to work with data consultants from a company that offers very specific business use cases and well-defined services. It's a plus if that consultant is backed by a team with a wide array of skills and expertise. Because, let's be honest, it wouldn't be the first time that a data specialist does not speak the language of the data owners. Not only does a multidisciplined team guarantee mutual understanding, it also widens the field of data applications.
The technology trapAgain and again, "data" is viewed from a technical perspective. Where is it stored? Is it in the cloud or on the premise? Is it relational? Is it accessible through an API? Many consultancy companies are keen to just add another piece of software/hardware to your technology stack that is linked to a strangling maintenance contract. Yes, they will open up the data silos and with some luck your company will achieve that highly anticipated 360-degree customer view. But what good is idle data? It's like brewing beer, storing it in a keg and not drinking it.
While it is true that technology often sets the context, turning data into value requires starting from the business case, translating it to a well-defined solution and then, and only then, should the underlying technology be the main focus. Not the other way around. You don't want to miss out on a highly profitable opportunity because you didn't consider it out of technological concerns. Remember that the mountain can always come to Muhammad.
It's more than data, it's your business
A consultant that is rooted in a full-service digital agency has a steadfast background in marketing and business and a fair share of digital skills. Working with a business-minded generalist with a focus on turning data into value ensures a verticals-first approach and that the problem does not get reduced to an intellectual rabbit hole.
The Reference is unique in the Belgian consultancy landscape; it fosters seniority, has a corporate mindset and works towards your business goals. We don't just dump a swarm of consultants at your doorstep. We maintain a maximum of in-house presence and at the same time we reserve space and time for our consultants to do research, exchange ideas with colleagues and get certified to guarantee the highest level of expertise in the future. Working with a data consultant from The Reference means working with a SPOC that is your company's access to a multidisciplinary data team and an interdisciplinary company within the hyperdisciplinary Emakina (ALEMK) group.
In the coming weeks, The Reference will release a wide range of data-related services. These will be accompanied by use cases, examples, interviews, podcasts and experiments. Finally we will also release our data-to-value framework; a set of tools that will guide your team from the ideation to operationalization of your most valueable data opportunities.
Do you want to stay informed and get access to our framework in the coming weeks? Subscribe to our newsletter below.