Where does Demandbase data come from?
The unique insights and recommendations that our products offer are driven by the proprietary data assets that our machine-learning technology produces. To do this, Demandbase begins by ingesting data from a wide variety of both third-party data sources and data signals that Demandbase tracks directly (first-party data). With the integration of the Engagio platform, we added Engagio's first-party data sources to ours, so we have access to even more data.
However, without our proprietary machine-learning algorithms, this raw data can be hard to interpret, incomplete, contradictory, or downright wrong. For this reason, our machine-learning algorithms compare all the sources to determine which ones have the highest probability of being correct, and then translate the data into the actionable insights that are specific to each of the products we offer.
In addition, due to the nature of machine learning technology, it is nearly impossible for us to attribute any one data point to a specific third-party source. In some cases, we may only end up using a small fraction of the data we get from one specific source. For this reason, we do not disclose our data sources.
How do Demandbase AI Algorithms work with data?
Demandbase’s proprietary data assets and algorithms power everything needed for a successful ABM strategy, including everything from identifying which accounts are interested in your products, targeting ads to those accounts, and personalizing your site when those accounts arrive on your site.
To make this possible, Demandbase ingests hundreds of millions of data points every day, from both third-party providers and our own direct and proprietary methods. The unique insights and recommendations that each of our products offer depend on these data points, but the data points alone would be meaningless without our machine-learning algorithms that house the intelligence needed to understand what they mean.
Our algorithms interpret these data points based on input from our market experts, historical trends, client-provided feedback, and training data we have accumulated over time. By regularly ingesting new data points and making human-driven updates to these algorithms, our models are constantly improving, yielding more accurate and more comprehensive insights every day.
How often is Demandbase data updated?
Many of our data points are updated daily. However, because our proprietary data assets are built up from sometimes tens or hundreds of different data sources, it's impossible to definitively say how often any one data point is updated because the primary and secondary sources may update at different intervals for each data point. On average, our identification and intent data is updated no less frequently than monthly, and our firmographic data is updated no less frequently than quarterly.
If you are dependent on third-party data sources, what would happen if one of them were to shut down?
Demandbase employs a wide variety of data sources to deliver our ABM tools and insights. While many of these data sources are collected directly by Demandbase, some come from third-party data vendors. Our machine-learning technology combines all of these sources to build a highly accurate and comprehensive representation of the business Internet which is far more accurate and complete than any one source alone.
That being said, many of our data sources provide overlapping data points; the removal of any one data source should not have a material impact on the resulting proprietary data assets and the tools and insights they drive. In fact, we regularly evaluate the quality of our data sources and remove and replace those sources on our own on a regular basis.
What should I do when I notice errors in Demandbase data?
Demandbase data uses probabilities through its AI algorithms, so we expect some errors, but, as Siri says, you're still on the fastest track to the best information. The more we know about your data, the more accurate future data will be.
Report any errors to support@demandbase.com. Demandbase’s data teams will evaluate your request and assess whether a correction is needed. If it is, the correction will be made and fed back into our algorithms to reduce the chances of a similar error being made again.