Is data science a bubble?
I’m being asked this question more & more, whether it’s from anxious data scientists or from concerned C suite.
We all know humans are susceptible to a bubble: our emotions create irrational market places. History is littered with examples, from tulips to real estate. The promise of untold fortune has led to speculation guised as ‘investment’, leading to an almighty crash landing once reality has kicked in.
Is data science the 21st century tulip?
Sci-fi writers & film makers alike have been creating the fantasy of the inevitability of our world dominated by killer robots & altruistic AI for literally centuries. Their worlds often include companies that created the ‘AI’ as king of the roost.
Also, have you noticed very recently how ‘AI’ is coming up more & more in the front pages of the major newspapers?
What was once found in tech publications, we now find regular 5 page pull outs in national publications on how AI is transforming the world we live. Bold predictions are made of a 300% increase in AI by 2022 with the inevitable background image of the robot from ‘I Am Robot’, coupled with case studies of evangelical business leaders gushing praise for how AI has transformed their industry.
There seems to be a real disconnect between the sci-fi writers, the technology journalists & reality for the vast majority of the industry.
Couple this with the rise of notably Amazon & Netflix with their famed data driven business models: the fuel for the fire of expectation is clear for us all to see.
However, for those of us in the industry on the proverbial front lines, the situation often feels very different.
From my experience, a more accurate picture of reality is the FD screaming down the phone: ‘Show me the money’ (insert Jerry Maguire GIF) to his data science team as he examines his balance sheet.
After 12 months of work, the data team tells the business: ‘We don’t have the data to do what we set out to do.’
7 figures sunk into a data programme.
0 demonstrable profit made.
The patience of business has been tested. Questions are being asked. The scores have been counted & it’s not a pretty picture.
Some have suggested that the whole industry is smoke & mirrors: a bubble, with its day of reckoning well within sights.
From my experience, there has been a noticeable change in business attitude over the past couple of years.
Historically, business essentially said:
‘We need to do something with our data, we want to do ‘AI’ (whatever that means), we need to hire the smartest person possible, the more degrees the better, whatever the price because they will solve all our problems & we’ll get that killer competitive advantage that’s clearly on offer.’
Demand went through the roof; people piled in. The smartest graduates on the planet 2 years into their career were earning £200k to solve the timeless philosophical classic of how to improve personalised advertising on Facebook.
Fast forward to today:
‘We’ve spent a lot of money & we’re not sure if we’ve got any value. What we want is people who understand our business & have demonstrable experience of building products that make a profit. We value data specialists with profit seeking entrepreneurial traits over pure intellect’.
The market has matured. Getting this stuff right is harder than first promised. Business is sceptical.
And that’s a good thing.
Scepticism challenges ideas & promises which increases the chances of the right decision being made – whether that’s on an individual or a supplier.
What we also know is that there is light at the end of the tunnel:
Real value can be created. Industries can be redefined. Human experience can be transformed.
Investors will always be drawn to AI driven versions of the future & want skin in that particular game.
So where does this leave the industry?
An analogy that comes to mind is that of a beginner surfer trying to catch a wave.
The surfer hasn’t quite worked out how to stand up on their board, but the surfer is being swept along by the ever energetic wave filled with enthusiasm. The wave seems relentless whilst standing up seems increasingly difficult as energy depletes.
(As an aside, anyone who has ever tried surfing will know that this is quite an accurate picture of the experience.)
Will she stand up & feel the rush or will she fall off & crash under the noise? Long term, who knows. The shore seems a long way off.
All we do know is that right now they have no option but to keep trying to stand up.
How do I think we need to stand up as an industry to this task?
– Asking tough questions of C level.
– Setting realistic business objectives.
– Taking disciplined consistent action.
– Fail fast approach.
It’s all too easy to point the finger at data scientists, accusing them of lack of delivering anything valuable, leading to the accusation of the whole industry being a bubble. However, the vast majority of the time the meta problem is the lack of direction, accountability & understanding at the very top of the business. This is what needs to be addressed.
Business has woken up to the realities of delivering profitable data products, but personally I cannot see a time in the short/medium term when business collectively turn their back on data science.
I don’t think there’s a bubble to burst. Machine learning is here to stay, at least in the lifetime of our careers. But business understanding has changed: expectations must be reset, tough decisions must be made and resources must be refocused.
Ammonite’s unique model was created because of this new paradigm of understanding in the data industry. Our focus is on creating real business value quickly, coaching the C suite on what’s required long term whilst building & delivering a sustainable resource plan.
Get in touch to discuss our services or this piece in more detail!