Data is almost universally seen as a good thing. The more, the merrier. Businesses stack their Gigabytes onto the Terrabytes and Terrabytes onto the Petabytes. They bask in the glory of total operational visibility. They brag about how it helps them solve any problem instantly, sell advertisements, or build v2 of their product.
But it’s not the data that’s valuable. The timely insights from timely data are the new gold, not the data itself. But why?
Data is a risk.
There’s a slew of tools to help mitigate risk around data ownership, but that risk will never be zero. We can build authentication, access controls, and encryption schemes directly into our application, along with audit trails that would make any SOC 2 Type 2 auditor smile. But complex systems are difficult to understand. It only takes one mistake, one unpatched system, one misconfiguration to expose your data. One insider. One successful fishing attack. The Internet Security Weather from bots and evil-doers will undoubtedly notice, exploit the opportunity, and leave your company with a pile of consequences.
Data limits growth.
Not all data technologies are equal. Early technical and architecture decisions limit how data can be used in the future. Further, International Data Residency, a term that showed up in recent Salesforce job postings, is a legal nightmare. Every country, and sometimes, every state, hold different legal answers to the questions,
- What is Personally Identifiable Information?
- Where is this data allowed to go? (i.e., what other countries?)
- Who is allowed to see this data?
- Who is allowed to delete this data?
- What can the company do with this data? (i.e., advertising, customer tracking, customer support)
- Can the company sell this data?
When a company reaches a certain size, you either have to solve it or face expensive consequences.
Data takes time and money to turn into value.
Data itself is not valuable to your business. Transforming data into timely insights is valuable.
However, going from data to insight is a challenging problem. It takes time and money, especially if you do it in-house. Machine Learning engineers are some of the most expensive engineers to hire in Silicon Valley, just like Security Engineers. But even when you have the team, your time-to-value is dependent on your business needs remaining consistent long enough to realize that value. Large companies tend to pivot less, but the initial time investment will still rack up to 12–18 months between,
- Building the team
- Identifying what business insights are most valuable
- Building the machinery to gain understanding from the influx of data
- Labeling data and tuning models
- Producing insightful reporting mechanisms
And even when your ROI starts to flow in from the business insights, your investment into this space continues to rack up. The operational overhead to maintain the thing, even if all future R and D ends, and these include,
- Cloud computation costs for report generation
- Ever-expanding data storage costs (If you use managed storage services)
- Operational costs (DevOps)
How can we use Minimalism to build a better future for data?
Data is not going away anytime soon, but the principles of Minimalism focus our attention on what really matters. So what really matters? How businesses insight can serve people. People are all that matter.
Throw away what does not add value.
Minimalism only works if we use it. To use it, we must focus on what really matters. We focus on what really matters by throwing away what doesn’t add business value.
The argument that the long tail may one day add value brings with it hidden costs, and I’m not talking about the cost of storage. Operational overhead to track, update, and accommodate International Data Custody, Laws, Regulations, Export and Import Compliance, and Privacy Legislation is difficult enough. It’s nearly impossible when they’re constantly changing.
Love people, use data — because the opposite never works.
The love affair some companies have with data destroys their relationship with people. Facebook is a case in point. But let’s not oversimplify. Facebook monetizes changes in the behaviors of its users, otherwise known as marketing. It shows advertisements to everyone. It uses insights derived from data to optimize which advertisements to show to which users and sells your change in your behavior (as a result of viewing an ad) to the highest bidders. And they save everything. But does what you did ten years ago in a Starbucks add value to what type of sneakers you might buy next? Probably not. So why save data from 10 years ago? The insights are valuable, not the data itself.
Let’s love people by recognizing the risks mass data collection poses to people with diminishing returns.
The Future is Insight, Not Data.
It might be interesting for Facebook to know that I came out as gay in Wayzata High School in 2001, in a heavily conservative area. Still, it’s not interesting to know where my mouse pointer was hovering over on October 12, 2001, when I logged into Facebook. See the difference? My coming out is evident in Facebook data, but the data itself was not my coming out. It takes time and money to identify opportunities, build tools, and operationalize processes that turn data into human insight and business value.
The fundamental difference between insight and data is that insight directly improves the company’s service for human beings (and we should always love people, not data). We need to use data to arrive at insights, but the value to the business is in the insight.
Data is dirty. Raw. Difficult and stubborn to work with. It consumes time and money to operationalize. It exposes companies to complex legal risks just by sitting around. But everyone needs the insights derived from all that risk to remain competitive.
More Insight with Less Time and Money
By centralizing the operational overhead, the barrier to entry melts away. Suppose we just let the data experts and lawyers solve this mess once and for everyone. In that case, all companies may enjoy the added benefits with less time and money before realizing the ROI.