Sign in Subscribe

A quick hit on how software companies can better realize the market opportunity for their data assets

Photo by Shubham Dhage on Unsplash

Pure-play data product and infrastructure companies have received considerable attention in recent years. Plenty has been written about data architecture, best practices in setting up a data startup, and the dos and don’ts of being a winning data company. But what about other “legacy companies” like vertical SaaS plays and internet marketplaces? I’ve seen what DaaS adoption can bring and think the same opportunities aren’t far off for many. Let’s unpack:

  • Hypothesis 1: there is immense market potential for data assets that are currently “trapped” inside of existing enterprises. Generative AI has brought this to the forefront: content companies are licensing or litigating their way around OpenAI and others who are using their content as training data. Many of these enterprises may consider themselves “data-driven companies”, but I doubt they view themselves as “data companies”. That latter term implies an operating model shift- a reclassification into a cutting-edge, advanced tech realm.
  • Hypothesis 2: the shift to productizing & commercializing data assets can be done alongside, rather than in place of, existing operating models. Just as the Associated Press can be a preeminent news organization while also licensing data and content to OpenAI, so too can a mid-market property management software continue to grow its core SaaS offering while licensing occupancy data into new markets.

First thing’s first: do I have an opportunity at hand?

While the above hypotheses are stated at the market level, the real question is “does my company have something?” Whether you’re an outside investor setting up a value creation plan or an internal product strategy team looking at new monentization opportunities, the underlying questions are the same:

  1. Do you have a unique asset: Are you regularly surprised at the uniqueness of insights uncovered internally? And I mean really surprised. If you’re planning on a side “Data Project”, you’re going toe-to-toe with competition solely dedicated to that task. Be crystal clear about your uniqueness and your data’s competitive differentiation.
  2. Is that asset valuable: Could you use those insights in marketing, investing, or policy-making? The “endless opportunities” thinking around data may seem a blessing, but it can cause analysis paralysis early on. I’ve found that these three are the largest centers of gravity in evaluating usefulness. Within these use cases, the same regular rules of TAM/SAM identification and market opportunity analysis apply, but these should help add a little structure at the start.
  3. How big is the lift to generate value: Do you feel like you have decent “data product” maturity? If your organization is already talking about internal data products, and has standards, shares, and services built around data, you’re likely not far off from taking those same products externally.

I think I have something, now what?

I mentioned above that commercializing data doesn’t have to be a fundamental transformation. But it’s obviously still a complex and complicated undertaking to launch a new business line.

So let’s overly simplify this process into 5 key concepts:

1. Isolate the asset

As an established SoftwareCo, you have one massive advantage on data startups- asset leverage. The most painful cost for a data startup is the cost of acquiring / processing large amounts of data until that asset becomes market viable. You get to skip that issue- it’s likely you have some form of input data you’re using in your platform as well as exhaust data generated from core systems.

Once you’ve ID’d the general assets, introduce Data as a Product thinking and turn those assets into products. If you already have, great. If you haven’t, you don’t need to reconfigure your entire Product/Engineering org, but you should create a tactical team that puts rigorous product thinking around the data assets identified above.

2. Think like a farmer, not a chef

As a SoftwareCo, like a great chef, you’re running an end-to-end organization fundamentally transforming an evolving list of ingredients into something uniquely outstanding to your customers. Opening an internal DataCo is the equivalent of taking a farm-to-table concept and making the “farm” part available to other chefs- your “customers” are no longer just diners.

In this world, your top priority is use case identification. In the World of Data, use case proliferation can be overwhelming so being clear, precise, and even a bit narrow at the start is critical. The two nearest-in use case flavors: a) sell your data into adjacent analytic needs at your more sophisticated clients and b) selling your data into other software solutions your clients are leveraging (e.g. customer data from syndicated surveys → MarTech targting solutions)

3. Say goodbye to demos

Ok, not actually. But you are going to have to rethink your pre-sales and sales motion. Data products don’t make for great demos, and it’s hard to “turn off” a free trial when you’ve sent someone a bunch of data.

Data buyers will want to get their hands on the data and there are two routes to do so, ultimately tied to your pricing model:

  • SaaS/Firehoses: send large chunks of data on a regular cadence as part of a contract. Aligns nicely with existing software operating models. But be prepared for lengthy, involved PoCs and long time-to-close to get clients over the sticker shock.
  • DaaS/Consumption: create APIs that run up a tally or draw down a credit bundle as users use more data. Viable if your data is valuable in discrete, small chunks. Painful for Finance predictability and for SaaS Sales Reps used to being compensated on big ARR wins.

4. The friend of my friend is my friend

Those use cases you identified above? They’re complicated and there are really savvy companies already selling into them. Domain awareness is especially hard when you’re navigating both this new data world as well as new end-industries.

If you’re unsure of value or don’t have great connects in the end-industry, look for channel partners (brokers, data platforms) and/or product partners (data-enabled SaaS). Rather than over-indexing on revenue, be ok with “sub-maximal wins” that generate domain awareness quickly.

5. Play in the sandbox

Again, second hypothesis- “this is work alongside the existing operating model”. But this can be messy work and, as obvious in the points above, there are a lot of considerations to be made that may not tie perfectly into the existing operating model.

If your goal is to build a strong, new revenue stream alongside your core business, you need to create an environment for a) learning and building but also b) not tracking a mess everywhere else. Set up dedicated reporting, test and learn procedures, a SteerCo — all of elements you’d normally put toward a major new initiative. Navigating the balancing act between “move fast and break things” and “don’t break the core business” is one of the more critical aspects of this whole process.

Recap: If you’re a strong, growth-stage software company, being a DataCo is likely in reach. No, it’s not easy, and no, it’s not simple. But it’s not impossible and you certainly don’t have to transform your entire business. This is work that can be done alongside the core, unlocking not only new growth avenues, but also new ways of thinking about your business. There’s being data-driven, and then there’s being a Data Company.

This is a summary view as part of my ongoing thinking on commercializing data products for “non-data companies”. Read on below for some expanded thoughts or follow along the series for deep-dives into each topic.

Series Contents

Time for Software Companies to be Data Companies

A quick hit on how software companies can better realize the market opportunity for their data assets