I think a lot of crappy data products and their supporting infrastructure are going to get scrapped in the coming months and years. Let me explain why. There is a fascinating positive correlation between central bank interest rates and the quality of your data products. When rates go down, the quality of data products goes down too, and when rates go back up, the quality of data products has to follow suit. With interest rates going up again, we can expect a lot of organizations to reassess their data initiatives.
Although I’ve written this piece for data practitioners, it contains a fair bit of economics, but bear with me. I’ve added a conspiracy theory as a treat.
As an economics student, I studied the Japanese Banking Crisis that shook the world in the 1990’s, and learned one thing:
“Cheap money hides systematic problems.”
In the case of the Japanese banks, cheap money due to low interest rates, and the resulting poor governance and a lack of regulatory scrutiny, meant banks were able to keep absurdly crappy loans, and overvalued stocks on their balance sheets instead of writing them off. The regular influx of cheap money artificially propped up a system ridden with inefficiencies which would fall down like a house of cards with the smallest gust of wind. The collapse eventually happened when the Japanese Central Bank started increasing their interest rate in ‘89. The whole thing came down in the most dramatic fashion. As markets started crashing and companies started defaulting on their loans, investors, supervisors and other stakeholders started questioning the high valuations of the stocks and loans on the banks’ balance sheets. As a result, banks had to write off these assets, drawing them in a downward spiral towards bankruptcy. It’s safe to say that a lot of people lost their cushy jobs pushing papers and drinking coffee at those banks. Something similar happened with the 2008 crisis, and is going to happen with the central banks increasing rates.
Early in my career, I transitioned from financial auditing to a data startup because the world of finance had become incredibly dull again, and I just couldn’t cope with the thought of having to wear a tie for the rest of my professional life. During that wonderful transition into the world of data, I learned another important lesson:
“Data is just another asset.”
It’s been a while since I’ve heard someone say ‘data is the new oil’ or ‘data is the new air’, for which I’m grateful. With the risk of sounding pedantic (which I am, quite frankly), data is just another asset. Your CEO and CFO invest in data and data infrastructure to earn money, and they read somewhere you need to be data-driven. They are expecting a return on their investments in data, and when that return is not high enough, they’ll redirect that money to where it actually generates a return. That’s it! Now, here’s the thing. As with any asset, Data’s value has been artificially inflated with all that cheap money flooding the markets over the past 10 years. Data literally exploded and a whole lot of new vendors came into existence in its wake. All that cheap money conveniently hid the problems that resulted from poor data governance, privacy and security. You need more data engineers to fix data issues? Throw some money at it! You can’t find your data? Throw some money at it! You had a privacy or security breach? Throw some money at it!
However, with inflation reaching levels that historically used to precede turmoil, central bankers are starting to soil their trousers. They have to hike rates, or run-away inflation will cause havoc.
And as cheap money dissipates, so will the artificial value of data. Your CEO, and CFO will start to question their investments in data, and will want to improve the quality of data governance to lower the cost of, and improve the return on investment on data. This will mean that only the data products that actually generate money will remain. All the rest will be killed off mercilessly.
I’m playing with a third lesson which is still rather anecdotal and maybe even a tad tin-foil hat, but I’m adding it anyhow for your entertainment:
“If government treasury is empty, taxes will go up, including privacy fines”
European and US governments are low on cash, which has typically resulted in the increase of ceilings for deficits and borrowing, and higher taxation. Here we get into murky territory and I have absolutely no proof for this, but as privacy fines go into the broader treasury, I wouldn’t be surprised that the increase of privacy fines over the past couple of years is partially driven by the need for more money.
Independent of my little theory, there have been way more privacy fines. Although Big Tech bore the brunt of it, privacy supervisors are increasingly shifting their scrutiny to the rest of the economy. As a result, your CISO and DPO will grow increasingly cautious about what you’re doing with customer PI and will want you to implement more privacy and security controls, which will further hamper the value of your already bad data products. Let’s be honest, if the need for better privacy and security is intrinsically an issue, your data product kinda sucks.
But hey, I’m not as pessimistic as it may seem from my writing so far. I’m just being provocative to keep your curiosity as a reader in a world that’s begging for your attention.
A correction is coming, but the end is definitely not nigh! Sure, you’ll have to get rid of some unprofitable data products, clean up your modern data stack, and improve data management, governance, and privacy, but I remain convinced that data is going to stay your most competitive asset, and that the trend towards data literacy and self service analytics will continue.
With the scale and variety of cloud data, increased data literacy, and the move to self-service analytics, companies will have to apply a modern way of data governance. Not one that is built around a central data governance team, but one where the responsibility is federated over the organization, and is tightly integrated with the development process of data products. To achieve this, observability, collaboration and automation will play a more important role than centralized control. Data Governance vendors will have to follow.