Data as a service (DaaS)

Data as a service (DaaS) is a business in which the product is the data itself. The seller collects data from many sources, cleans it, matches it and keeps it up to date, and the buyer pays a subscription to use it. Such sellers are also called data vendors. A company buys data this way when it needs facts that it cannot collect from its own systems, such as how many phones every shop in the country sold last month.

Here is how it works in general. The vendor gathers raw data from many places: shop tills, households that record what they buy, public filings, websites, cameras, sensors, or people who record events by hand. Each source describes things in its own way. One shop records a phone as "Brand X 5G 128GB", while another records the same phone under a product code. So the vendor's software cleans the records and matches them to one master list, every day. The vendor checks the result for gaps and errors, and adds its own estimates, for example for shops that did not report. Buyers reach the data through a dashboard, a file delivered every week, or an API, which lets the buyer's own software fetch the data automatically. The price of the subscription usually depends on how much of the data the buyer uses, and how many of its staff use it.

Most of the cost is in this engineering. Collecting the data once is not enough. It has to be collected and matched again every day, for years, so that this month's figures can be compared with last year's.

Take, for example, a soap brand that wants to know its share of sales in Tamil Nadu. Retail measurement combines till data from thousands of shops with household records, and sells each brand its share every month.

Consider a parcel delivery company that needs an accurate map. A digital map is a dataset too: it is gathered from satellites, vehicles and surveyors, updated all the time, and sold on subscription.

For instance, a law firm pays for a legal research database. The judgments themselves are public, but the vendor sells its summaries and the links between the judgments.

How is data as a service different from SaaS (software as a service)? With SaaS, the buyer rents software and puts its own data into it. With data as a service, the buyer rents data that someone else has collected. Data as a service is also different from an analyst firm, which sells experts' judgement and advice rather than a dataset. One common way of collecting such data is described in consumer panels and retail audits.