Data-as-a-Service: Running DaaS Companies
blog.safegraph.com
blog.safegraph.com
The first challenge I see with data as a service is that not all features you need for your model will be made available by the vendor.
The second is their frequency of update might not be at the cadence that you need.
Having seen both the selling and acquisition side of the table, the only concrete advice I can give you is to pay handsomely for experienced sales staff. There is absolutely no rhyme or reason to data pricing, and no matter what you do you will invariably price out specific market segments while underpricing for other market segments. And as the article mentioned, there are a lot of potential licensing and usage covenants involved in data sales, which are all possible factors in pricing. Without a good sales team to do client and market discovery, it's really hard to understand what to settle on for norms as far as both pricing and licensing go.
On the flip side of that: if you're ever purchasing data at any amount of volume, ignore listed pricing. It's absolutely fungible, and well worth the tedious dance of traditional sales. In addition to better financial terms, you can usually get adjustments to the licensing and usage covenants to better accommodate your use case. Doubly so if your use case is nonstandard for that vendor - plenty of data-oriented services are tailored for specific industries and uses, and their pricing sheets and tiers are designed around the use cases and marginal value of the data within that industry. When sourcing data, I could get pretty incredibly deals from these companies, as their sales teams basically used the transaction as a form of market discovery and saw the relationship as a way to feel out a particular market segment they hadn't been going after previously.
Then the founders stumbled into it's potential worth when the access required for that purpose was secondarily leveraged for marketing intelligence purposes, and was bought by (what's currently) Slice Technologies.
Even when companies aren't using dodgy tricks, you've got to keep an eye out for ways they could if they felt like it.
After 6 months of failing to hire in Lisbon a company I know still won’t spend 3k to work out what the local market expects from the sort of job they advertise through a representative sample survey ... they fully understand why it’d help, it’ just esoteric to them . Unfortunately this is pretty normal .
There are more good uses for data than there are good users.
So- think about the personal microdecisions you make on a daily basis, especially the small ones- shopping/cooking/eating; cleaning; dressing; commuting; etc- and appreciate how many of those are not data-driven, are rather just based on assumption, habit, history, etc. Appreciate how many things one wonders about, idly, throughout the day, in regard to those microdecisions.
With some reflection it's easy to see that having various kinds of data would lead to your making different micro decisions. Some of those microdecisions can substantively impact things that are important- health, wealth, happiness- so will become business opportunities. Apply that to microdecisions made in the course of work...
People will buy data when it has a clear use which means most people don’t sell just data, they sell “leads” or “clicks” or “followers”: something actionable.
I think this is unlikely to change because having an internal data team producing leads typically doesn’t make sense as a core focus.
There are a relatively small number of companies which do sell data and a slightly bigger number of companies that actually buy data. In the UK it’s easy to quantify because they all end up buying some of exactly the same thing.
What exactly is this thing?
You are right though. People don't buy data, just like they don't buy a SaaS product. They buy something if it solves a direct need, saves them money, or helps them generate more revenue. And like all things, it does need to have a clear, understandable story and value proposition.