Why sell data - Part 2: Packaging & positioning

One of the things about most datasets is that they can be used by different types of customers for different reasons. Narrowing down exactly what you’re going to sell is one of the most important pieces to being successful.
‍

If you missed it, you can still read part 1 of this series.

‍What makes data valuable?

So, let’s start with the question of what makes data valuable in the first place.

And the answer is that -> data is only valuable to the extent that it solves a problem.

The more valuable the problem you can solve, and the fewer alternatives to solve that problem without your data, the more valuable your data is!

For example, Solomon Khan, Founder & CEO of Delivery Layer and co-author of this post, used to work for Nielsen, which has a TV ratings business to measure how many people watch TV programs along with demographic information for those people.

For regular people, this data is interesting, because it’s always fun to know what shows are the most popular, but nobody would pay anything for it.

However, advertisers are a different story. In order for buyers and sellers of TV advertising to transact, there needs to be data on how many people watch a specific advertisement.

Without Nielsen data, the $60 billion yearly TV ad market would not exist in the way it does today. For years companies have been trying to find cheaper alternatives to Nielsen data, and every time it quickly ends in disaster and they go back to buying Nielsen data. Because it solves such a valuable problem!

So -> same dataset, with one customer type it’s a random interesting tidbit that they wouldn’t pay for, with another customer type it’s a multi-billion dollar business.

‍Creativity

It’s not always so obvious who can use your data to solve valuable problems, and people have come up with some very creative “alternative uses” for data.

There are examples of where people have:

• Used satellite photos of parking lots to estimate sales for large retail stores
• Used mathematical models to estimate shipping by measuring the level of randomness on tracking numbers
• If ever there’s a company that increments things in an obvious way (e.g. order 132, 133, 134) that will no doubt be used as a signal
• Etc…

This works for investors because they have an evergreen problem of “how can I use data to predict the future in ways that I can trade on right now to make money.”

There is an emerging subset of data consulting providers that focuses on extracting this external value out of data. These companies, like Neudata, help data owners find use cases for data and then package and sell those new data products to third party data buyers.

‍Consistency

If your dataset is needed consistently, it’s generally more valuable than if it’s needed once.

Back to the fundamentals, recurring revenue businesses are valued more highly than project-based businesses for a reason.

‍Data attributes

The market for data has been around for a while, which means there are already many datasets in the market. Yours can differentiate itself through its:

• Uniqueness -> you have information or insights people haven’t seen before
• Timeliness -> For example, daily vs. monthly or quarterly
• Granularity -> For example, zip-code level vs. country level

Ultimately, the value of data is based on the problems it helps solve. Whether it's enabling billion-dollar ad markets or uncovering subtle signals to improve business or investment challenges, the most valuable data unlocks insights, decision-making or profit. Sometimes, all it takes is a shift in perspective - or a creative partner - to uncover those opportunities.

‍Use cases

‍Investment

In the investment world there is a constant need to forecast the future because having an information edge can make you a lot of money. That’s why investors are willing to pay for data that helps them see what others don’t.

This is true for hedge funds which can trade immediately on those signals, as well as PE firms that use data for M&A due diligence, trend analysis, benchmarking or portfolio value creation initiatives.

‍Corporate

Corporations benefit from increased intelligence about their industry. Companies pay big for all sorts of data that helps them with:

• Market-share tracking
• Marketing performance and effectiveness
• Understanding prospect and customer behavior
• Brand perception and positioning

‍Operational

Businesses are more interconnected than ever before, which means that operational / partnership / supply chain vendors, who you would not necessarily lump into the typical “corporate” bucket, might actually be very interested in your data.

For example:

• Vendors and partners using your data to forecast demand, manage logistics or negotiate more effectively.
• Supply chain participants improving efficiency or de-risking operations through better visibility.

These use cases often go unnoticed by data owners - but they can be lucrative if you find them.

‍AI

Some companies have data that is highly useful for AI purposes. Both for foundational training models (e.g. Reddit) or for specific use in AI workflows (e.g. corporate insights data to be inserted into email sequences).

If you have a dataset that’s useful for AI, the rise of AI is good news for you.

‍Commercial models

There are many ways to sell data as with any product, but three primary ones stick out as by far the most common.

• Fixed licence fee - Pre-determined price for data access and usage. For example, $50k per year for full ongoing access to the data.
• Usage based - Pay only what you use. For example, $0.05 per customer record.
• Project based - Typically a “single cut” of data for a specific project. You only need it once, you only get it once, and you only pay once.

The right model depends on your data’s structure, frequency of updates, and how customers will use it.

‍Delivery methods

You can deliver data products in a variety of different ways - depending on what your customers need and how they want to use the data:

‍Files

Still the most common delivery method for large amounts of data. You add a file to an S3 bucket or an FTP server that the customer has access to, and then they download the file and do whatever they want with all the bulk data.

‍Applications

Sometimes customers don’t want raw data, they want an insights platform that will let them see and analyze the data in charts, graphs and data tables.

In this case, they need some web application. Delivery Layer is a platform built specifically for these types of applications. Also, some data sellers build customer applications with web development teams, while others use BI tools, either in their classic setup or through an embedded BI application.

‍APIs

Sometimes, customers don’t want raw data, they just want one single data point. For those cases, data sellers build a lookup API so that data buyers can get all the data they need one row at a time.

‍Data shares

Many of the modern cloud databases allow companies to “share” database tables. In this case the data buyer can query a table owned by another company as if it was just part of their own database, with no data loading required.

‍Email

Lastly, we can’t forget the age old delivery method of sending someone an Excel file or a PowerPoint with the data they bought. As much as there are many high-tech solutions for moving large amounts of data, email is still quite common.

‍Packaging

One of the most interesting things about a data asset is that you can package up the exact same data into multiple different products for different customers depending on which problem they are trying to solve.

Using Nielsen again as an example. Co-author Solomon Khan worked in a division that analyzed each second of professional sports across the world, identified which brands were on the screen, and then used that brand exposure data to support companies operating in the sports sponsorship market.

As you can imagine, building that kind of dataset - measuring brand exposure across global sports programming - is incredibly costly.  But once it’s created, it can serve multiple purposes:

• Teams use it as part of their sales and customer support processes. They need charts, dashboards and rollups of specific “inventory spots” and customer reporting for partner brands. They want to tell a story about the value of the exposure generated.
• Brands look at it from a portfolio perspective. They want a high-level view across all of their sponsorship investments - across teams, leagues, and even countries - to evaluate performance, compare assets and inform where to invest next.…
• Leagues need league-wide sponsorship assets and data to help teams or broadcast partners generate value.
• Athletes use data for supporting negotiations for brand partnerships, sponsorships, or other negotiations
• etc…

The underlying dataset didn’t change, but the products, visualizations, sales pitches and pricing models are entirely different for each of these buyers. 

One dataset, many revenue streams.

‍Explore opportunity areas

When you’re first exploring how to sell your data, you should explore the entire opportunity space for all the possible customer types and all the potential use cases.

Hedge funds will use your data in VERY different ways than shipping companies or marketing execs. But they might all be interested.

Don’t stay at the industry level though, go all the way down to a specific person at these companies that will be interested. It’s not “shipping companies” it’s “recruiters at shipping companies” or “revenue operations leaders at shipping companies” etc…

What is a major problem they are trying to solve in their jobs, how does your data solve it for them better than the alternatives, how valuable might that solution be, and how big of a market might the overall solution be?

The bottom line is you need to ask a lot of questions, meet a lot of people and kiss a lot of frogs.  Finding a good partner who knows the space can often accelerate the process and increase the chances of success.

Then focus

You might start with many potential uses and customer types for your data. Ultimately though, any successful business very tightly solves specific problems for specific people.

You might have 5 different ways you can package your data for various customers in your ecosystem. Once you have a good enough view of the market potential for your data assets you can start focusing.

It’s best to start with 1-2.

And your marketing shouldn’t be about how great your data is, it’s all about your potential customer’s problems and how your data can help solve them.

Over time, if you’re successful, you will have the revenue to launch new use cases for new customer types. But to start, focus on the 1-2 major things that you think are most compelling.  You can't come out of the gate being all things to all people.

Once you have those 1-2, you’ll be ready to build a test and take it to market. We’ll get to that in the final post on this series of selling data.

If you missed the first post and aren’t quite sure about why you should sell data, learn here about the benefits, risks and market potential. 

‍

Neudata consults with companies working through these packaging questions. Delivery Layer builds reporting applications to serve those different customer segments. Reach out to either of us if you want to chat about any of this.

‍