For most companies, especially startups and tech, data is your most valuable resource. Data helps you get customers, keep customers, build products customers love, raise capital, allocate capital, etc. Data helps you make decisions - better, faster, and with more precision.
At the same time, data is your most under leveraged resource. Data and analytics practices are rarely built into the design of a company. Instead, they're bolted on after the fact, like a hand-me-down dress suit that kind of fits and smells weird. This article will dive into why and how this happens, why getting it right matters - a lot, and what to do to fight against the gravity pulling towards data hell.
Before we dive in, why should you listen to me?
You probably shouldn't. At least on the basis of who I am. You should decide whether my POV holds merit on the basis on my arguments and evidence.
That being said, I have been at this for most of my career (such as it is). I've seen the good, the bad, and the befuddling. I've built Marketing, Data, and Growth functions at 2 early stage startups after spending 6 years at Amazon - the mothership of data and operational efficiency.
The Tectonic Shift in Value Creation
The commercialization of the internet and the advent of the World Wide Web (Tim Berners Lee! A great story for another time) brought humanity into a new age. The age of information. As information and digital technology became increasingly ubiquitous - a familiar refrain propagated: software is eating the world and big tech is the manifestation.
That wasn't always the case. In fact, it wasn't the case just a few years ago. So what happened?
In 2017, The Economist ran this cover story: Data is the new oil.
Although, I'd argue that energy is still a more important resource than data (need energy to get and use the data), the new paradigm was officially here. And the rockstars who came down the red carpet this time were tech companies, who had more data and access to future data than other companies.
Looking at the top 20 companies by market cap over the last 30 years, that's exactly what happened.
In 1993, energy, consumer, and health care dominated the list. While I'm sure those companies used data and analytics, their products weren't data products and they certainly weren't internet companies.
2003, technology and finance arrived on the scene, joining the other folks. And again, while those tech and finance companies certainly benefited from increased information - they weren't internet companies and they weren't using data to create yet.
2013, pretty much the same story as 2003 and many of the same names. Alphabet enters the list, marking the first company borne out of data to reach the top 20. Additionally, Apple - which is and was primarily a consumer hardware company took the top spot. Apple devices connectivity arguably makes them a data company too, but not really by this point.
By 2022, the game had changed - both in the companies creating the most value and the business models of tech companies.
The first case, enter stage left Alphabet, Amazon, Tesla, Meta, Nvidia, Broadcom. All data companies, born out of the digital information age, with data creation/consumption core to the business model. The second case, names like Apple, Microsoft, and the cadre of financial services - all have data creation/consumption as a core pillar of their business model. And I posit possess an even more defensible position than the platform businesses.
I believe in the foreseeable future, most companies will be tech companies - at least in how we'd categorize today. Whether you agree or not, what's indisputable is the incredible growth and value creation. The value of data is increasing exponentially and I don't see any reason it would slow down. In fact, the dawn of AI in the general public this past year would suggest it's going to accelerate from here.
So if data is so valuable, what gives? Why do so many companies, particularly startups fail to leverage it?
Why Companies Become Data Poor
Inaction Inertia - as I mentioned above, data and analytics aren't typically built into the company from the start. The more times the organization declines the opportunity to begin harnessing data, the longer it goes without using quantitative information - the less likely taking that action becomes.
*Side note - one of my college professors was one of the researchers who proved out 'inaction inertia', shoutout Mr. Thane Pittman :)
The Data Wheel of Death - Data isn’t constantly maintained. Data becomes less useful. People lose trust can be applied. The company uses less data.
This is super common and relatively self-explanatory. If one can't see the big picture and execution concurrently, this looks like a failed investment area. Most of these cases are well meaning teams, who run through brick walls and end up tiring themselves out.
Don't Know What They Don't Know - This is the most dangerous trait in business. Not knowing something is okay, not knowing you don't know something is not. In some cases, folks don't have the lived experience of a data rich organization and they're not aware of their blind spot.
Or they think they are building data and analytical practices, but don't have the awareness/imagination to see beyond what they themselves are capable of. A telltale sign is the use of proxies, dashboards, over-summarized output metrics, and samples sizes of one. There's a better path, I promise.
They Lack the Vision - Building on the last one, some folks aren't able to see the applications and therefore, can't get the value. This is part technical skillset and part commercial strategy. It takes both parts to put the data to good use, and sadly, those two symbiotic skillsets aren't always married.
The right data and tooling enable anyone in your company to find the information they need and act on it - regardless of technical ability. The information flow will empower decision making, prioritization, and an owner mindset across the organization.
Why This Matters
Decisions, Decisions, and Decisions
I propose everyone inside your organization is an analyst. Perhaps not in title, and analysis isn't all they do - unless you're in a state of paralysis :). But everyone inside your organization makes decisions, big and small, all the time. To make a decision, one must analyze situations, construct models, and project consequences. Making decisions is ultimately what we all do in life. I can't think of a more important thing to do well.
What's the best predictor of making good decisions? The right amount and type of information. If one has the right information, making the right decision becomes much more straightforward. When good intentioned decisions go awry, it's because the right information wasn't present.
Now, I'm not saying every decision has or could have the right information. The frustratingly magical thing about the world we occupy is there are many facets we don't know and cannot predict with certainty. Outcomes are often not controllable.
I am saying it behooves an organization to maximize the ratio of decisions made with the right information. That is a key input to decision making and it is controllable.
Flywheels and Learning Loops
Data can be a powerful flywheel for any business, but only if it's usable and readily applied. Two virtuous cycles to illustrate.
Nick's Marketing credo:
Taking action creates information.
That information informs and guides strategy.
In turn, the strategy enables decision making.
Decisions create action, which leads to more information.
Nick's Product/Website credo:
User interactions create data.
Usage data informs UX and growth design.
Better UX and growth design will help attract more users.
Those users create more data, and around the wheel we go.
Competitive Advantage
This one will take a few sentences to unravel so bear with me.... Let's think about sustainable competitive advantages. What is a competitive advantage a company possesses which sustains itself over time?
My first thought is talent.
Companies are no more than a cohort of individuals creating value for others. The world doesn't change on it's own, people change it. However, the war for talent has winners and losers. And the winners today are the losers tomorrow as the fate of companies/industries changes.
If talent doesn't hold up against time, then culture doesn't either. After all, culture is created and maintained by people. If the people change, so does the culture (although more slowly).
Business and operating model perhaps?
If products, technology, and talent don't sustain an advantage, then this doesn't work either. The business and operating model will change or perhaps be supplanted by a better one.
Category and brand?
As a marketer, I wish this were true. But categories change. Just ask Mad Dogg Athletics (Spin classes) what happened when Peloton arrived on the scene. Or a thousand other examples. Brand capital can persist over time, but not without the category and categories change.
So, what does sustain over time then?
As you may have guessed, I'd argue your own proprietary data is one of the only sustainable competitive advantages. You can keep it forever, it's a unique and increasingly valuable resource - and that may never have to change. It's not a silver bullet by itself, the other things mentioned above need to be there, in concert.
What You Can Do About It
If you've made it this far, you probably want to know what are some things you can do to build data and analytics practices in your organization. And you've definitely earned actionable tips :).
Intro Activity
On teams I've been a part of in the past, I used this exercise as a pre-cursor to standing up data infrastructure and analytics practices. These questions are particularly helpful if you're struggling to get buy-in.
Does increased information, measurement, and analytics support your company values? (assuming here you have written values)
If we had better data, what questions would we be able to answer right now?
How about in the next 12 months?
What decisions are we making today without data, that could be informed by data?
If we don't improve our data and analytical practices, what will we jeopardize?
If you had a magic wand, what information would you possess today?
What metrics would help you do your job better?
Assign Resources
"The best way to ensure that you failed to invent something is by making it somebody's part-time job" - Dave Limp, SVP, Devices & Services at Amazon
Someone needs to own it. Like any other centralized system you use, it'll become a total mess unless there's a gardener. Ideally, a gardener with the chops and vision to lead the initiative.
Take Ownership of Your Data
Get minimum viable tools in place to capture and store your data, as soon as feasible. There's a plethora of open source analytics platforms to capture with (I like Posthog and Rudderstack, tons of options in the tools catalog).
Storage costs have plummeted since early cloud days and ease of use is way better. It's one of the more obvious cost to benefits there is. For this, I like Redshift from the start, if you're later on without a warehouse, Snowflake is a good option.
To get most of your data in there, you'll need a piping tool - Airbyte is my favorite right now, lots of options in the tools catalog.
Think about Data as Product
Approach it as an internal product, with a roadmap, user research, measurement, and structure. Start small and build momentum. Then expand, saturate and build momentum, and expand again.
Apply It!
Dovetailing from data as a product, start applying the data you have and stretch that muscle. Integrate the data into the teams it affects ASAP. Go back to the intro activity and knock some of the low hanging fruit out. Find a decision with $s and ROI attached to it. Run an experiment. Start reporting on metrics, sharing with stakeholders. Lots of ways to get started here.
If you're super early and just have things like Google Analytics and/or a CRM, two tools that are great to get started with are narrative.bi and plusdocs.com. Two user friendly (read no code, spreadsheet logic) data science tools to start performing analyses and modelling that I love are count.co and einblick.ai. Again tons of resources in the tools catalog that can play in this spot.
Comentarios