It’s something that’s been happening slowly over time, yet also (somehow) conversation around it seemed to change suddenly overnight: business intelligence (BI), which was once the sign of a thriving data-driven culture, has given way to data science, shifting attention to machine learning (ML) and artificial intelligence (AI).
At the time, the shift was so gradual that it was easy for companies to miss (or they lay in wait for the changes to stop before taking action — but they never did). Now, suddenly, those still trapped in the world of BI are left behind (far, far, behind). Here’s how we got here and why you still need to take the leap (even if you think it’s too late to catch up).
We’ve Safely Arrived in the Era of AI
Just in case there is anyone out there still unsure, in mid-February, Forbes released its Roundup Of Machine Learning Forecasts And Market Estimates, 2018. In it, they proclaimed:
Data science platforms will outperform the broader BI & analytics market, which is predicted to grow at an 8% [Compound Annual Growth Rate - CAGR] in the same period. Data science platforms will grow in value from $3B in 2017 to $4.8B in 2021.”
What’s more, the International Institute for Analytics recently released a study showing that analytics maturity actually has a huge impact on company performance. If businesses increasingly putting their money in data science and removing focus from BI isn’t a sign of this shift, then what is?
What Happened to BI?
When businesses started collecting data and deciding to actually use it for something (we’re talking the late 80s — this is a good read if you want a more in-depth, non-simplified history), it was small potatoes. Data coming in was small enough to work with locally and, since analyzing data was new, insights were revolutionary (looking at some data is better than the previous way, which was looking at no data).
Now with the amount of data we have today and, moreover, the differences in the way people consume and use the internet in general, it seems crazy to be using the same analytical methods as before.
And, well — it kind of is crazy. Of course, there have been lots of developments and shifts in the world of BI since it became standard business practice, and it has become much more advanced. And it’s also worth noting that there is a time and a place for good BI. But the reality is that lots of businesses are still really only doing BI today; that is, they are reactive — taking past data and using it to influence future decisions in spite of the fact that our world today is fundamentally different.
With the enormous amount of data being generated today, we finally have basically up-to-the-minute information on what people do and evidence that consumer behavior can (and does) shift on a dime. It’s clear that reactive analysis is no longer the answer. Yet enterprises are still doing it.
Talk Is Easy; Action Is Hard
The reason enterprises haven’t fully embraced the shift to AI is actually pretty simple. Though from the way people speak about it (news coverage, popular culture, etc.) AI seems easy or like something that can be activated by throwing some money in and flipping the switch, it’s actually a fundamental challenge that is organizational as much as it is technical. Maybe it’s not having the right people, the right expertise, the right tools, the right motivation from leadership, the right mindset - the list goes on.
In the words of the great Elvis Presley - "A little less conversation, a little more action (please)."
Today, basic access to data that is accurate and clean can still be a surprisingly major problem for large, established enterprise, even those with a good reputation for innovation. If everyone is blocked at step zero, doing steps one, two, three, and beyond at a huge scale and automating it is daunting.
Get Started Now
The good news is that it’s not too late to get started. And while there’s no magic bullet, there are some pretty accessible things your company can do to make the change:
- Set a goal from the top to become a truly data-driven organization, meaning one that is powered by massive amounts of data all the time, not reacting piecemeal with one-of analysis. See the video for more on how to set (and achieve) this goal.
- Leverage current staff (instead of hiring new people) by providing them with tools that allow them to work with data while also leveraging their business and domain knowledge. This might mean teaching analysts new skills that allow them to be involved in machine learning projects. Get the analyst of the future guidebook to help or download machine learning basics to aid staff in their journey.
- Arm technical staff with tools as well that will allow them to be more productive by reusing their work and leveraging shortcuts where it makes sense (like visual data wrangling and easily comparing the performance of different models side-by-side). Read more on what data science tools can do.
- Introduce data science, ML, and AI in an accessible way by choosing a specific use case with which to start and providing a framework for making it happen. Read more on how to run a data science POC successfully.
- Have a plan for putting the final data project into production to ensure that it actually has a real effect on the business instead of going unused. Read more on putting data science in production.