How Switchboard helps publishers build a first-party data asset

Last week, at AdExchanger Programmatic I/O New York, we spoke to fellow data executives and prospective customers about how they can make the most of their own first-party data. Here, we share the benefits using it, as well as the secrets to our customers’ success.

The benefits of using first-party data

We recently discussed how publishers are getting prepared for life beyond the cookie, and how third-party data has never really been the best option, due to inaccurate data and lack of campaign control. As we see it, first-party data should be enough to provide publishers with the insights they need to keep both their advertisers and audiences coming back for more.

The main benefits are:

• Better privacy. Consumers are aware of privacy concerns relating to third-party data thanks to legislation in Europe and the US, such as GDPR and CCPA. While you need to remain responsible for your data (as set out in those laws), it becomes less of a headache when you’re only dealing with data you already own.
• More control. With traditional third-party ad campaigns, you’d have a “persona” created for you based on that data, and you’d run a campaign on it. But inevitably, if that data is inaccurate, you don’t get the results you want. First-party data allows you to better understand the reality of your customer profiles.
• More accurate reporting. Third-party campaigns typically use generic, high-level reporting, in which some of the metrics might not matter to your business. But with first-party data, you have the ability to build unique reports – according to your teams’ specific KPIs – that your business really cares about.

So as a publisher, how can you use your first-party data to its maximum potential?

How Switchboard helps

There is one main caveat to using first-party data as your go-to asset: it takes time to master your data set and learn how to use it in a way that makes sense for your business. So it’s best to start as soon as possible, but also to implement a layer of automation to do the heavy lifting.

Publishers have tons of data coming in from different sources, from critical billing and sales operations, to growing audience data across multiple properties and categories. You could hire a group of developers to build bespoke software that collects, sorts, and loads the data into a data warehouse, but the time and financial cost would soon become unsustainable.

Instead, you can use a platform like Switchboard, which automates the process, without your engineers having to write line after line of code. In fact, Switchboard takes care of the entire ETL process:

• Extract. Switchboard extracts the data from all your different sources and aggregates it, while allowing you to set parameters such as time zones, backfilling missing data sources, handling APIs, and compliance with data security.
• Transform. Switchboard will then clean the raw data and transform it by applying your business’s rules (what we call ‘data recipes’). These rules include standardization, verification, formatting and sorting, labeling, and protection.
• Load. Finally, Switchboard will load the transformed data into a unified data storage warehouse or data lake. From there, you can catalog, maintain, archive, and perform other data governance tasks.

Once the ETL process is complete (and while it’s continually running new data in the background) you can use the unified solution to parse complex dimensional data, for crucial insights into advertiser campaigns, performance across properties etc. You’ll be able to work with your own first-party data asset, with far more accuracy and privacy.

This process happens automatically and in real time, giving your business teams the ultimate reporting ability. By using automation in this way, Switchboard customers have been able to save quarters, if not years of custom development time, not to mention multiple person-hours per week maintaining their data sets.

To find out how Switchboard helped the Financial Times stand up a sell-through forecast in one week, and eliminate a manual bottleneck of four-person days per forecast, read the full case study here.