For a significant period of time, and until relatively recently, registered California data broker SafeGraph was providing an “Ad Targeting” user data appending service that openly used a 100+ million person-to-home-ownership database in the United States, openly scraped from U.S. government data sources, which was owned by a Delaware-registered shell company called “BA45.”
The research included below has been deeply fact checked and additional documentation is available as-needed to prove the significant concerns about SafeGraph and their partner BA45.
It should be noted that BA45 has never been very open or public about their work, and the CEO and founder seems to have never promoted their own company on social media. That being said, to be fair, I reached out to the BA45 founder with details about this research, and they sent over these details about BA45:
“BA45 was a pre-seed business bundling public government records data. The business was unable to acquire enough customers and has since shut down.”
BA45 was founded in Delaware on July 30, 2020 by an OG Yandex social media data scraper, entrepreneur and data scientist who had recently sold their own B2B data broker company, Orb Intelligence, to Dun & Bradstreet. The sale was described in Russian media as the “amount of the transaction amounted to several tens of millions of dollars, but did not exceed $ 50 million.”
The founder of BA45 started work at Dun & Bradstreet in January 2020 and seems to be currently their Senior Vice President of Engineering – and this same BA45 founder was seemingly working at D&B when BA45 was founded in July 2020.
The founder of BA45 seems to have worked at Dun & Bradstreet during the entire time BA45 existed, during the entire SafeGraph partnership, and when BA45 was appending U.S. consumer data scraped via U.S. government websites for as SafeGraph’s own marketing described it, “Ad targeting.”
It has never been confirmed or clarified whether Dun & Bradstreet were aware of BA45, whether they were a source of data, or an investor in the business — no efforts were made to reach out to D&B to determine this information, but questions about D&B were sent to the BA45 founder before publishing, which were not answered.
The BA45 website (now also taken down) was saved in the Way Back Machine last year here. The “Download dataset layout” link from their homepage is still active on the archived page, so you can download the original version – a copy of that has been uploaded to Google docs here.
Someone could also book a call to learn more about BA45 through the now-defunct website:
The SafeGraph BA45 Data Partner Page did not hide (until they took down the webpage) that BA45, “sourced from more than 3,000 U.S. counties and web sources” – this is a casual way of saying, “this data was scraped from the U.S. government.”
The BA45 founder confirmed this U.S. government data source in their response to my questions.
SafeGraph.com is a registered California data broker who had their mobile SDK banned by Google for harvesting and reselling user data.
Vice also previously reported that about SafeGraph’s investors with ties to foreign intelligence services, writing “In April 2017, Turki bin Faisal Al Saud, the former head of Saudi Arabia’s intelligence agency, invested in SafeGraph as part of a $16 million Series A funding round.”
SafeGraph’s website openly declares, “It’s Time for AdTech Companies to Marry Location-Based Data with Audience-Based Data” while claiming lower in their same post a false claim about how they handle user data for SafeGraph client appending:
“We take a lot of pride knowing that our data maintains the highest level of security, quality, and compliance standards in the industry. But unlike other data providers, we don’t face the same legal issues or other GDPR and CCPA compliance-related limitations because all of our data is 100% anonymized. We do not track individuals whatsoever. For this reason alone, SafeGraph is the ideal data partner for AdTech companies looking to bridge audience-based insights with location-based data. (And because we provide clean, accurate, and high-quality data, too!)“
BA45.com SafeGraph Partner Openly Admitted to Acquiring Property Data with People’s Names from U.S. Government Websites to Support Ad Targeting and Deanonymization Services for Less Than a Penny Per Record
Here’s how SafeGraph previously described the BA45 partner before they were removed:
The SafeGraph BA45 Dataset was described as containing data on “ about 125M commercial, residential and governmental properties”:
“BA45 properties and owners database contains information on about 125M commercial, residential and governmental properties. Each row represents a property and comes along with 40 attributes describing the structure (building), land, sales price and date and ownership information. Owners’ names and addresses are parsed and unique owner ID is assigned.”
The BA45 + SafeGraph sales page goes on to describe the offerings, including:
“Underwriting Risk – Insurers utilize BA45 property data to understand up-to-date building characteristics and calibrate their pricing models accordingly to develop a comprehensive understanding of risk. Attributes about the property are essential to appropriately underwrite insurance policies.”
“Prospecting Borrowers – Lenders use property data to assess homeownership and prospects for new mortgage financing or refinancing. Mortgage insights enable lenders to manage their portfolio and create leads for new financings.”
And most importantly, the BA45 + SafeGraph product for ad targeting – which was described as “Pristine and up-to-date” :
“Ad Targeting – Marketers leverage property ownership data to create a comprehensive view of audiences and more accurately segment and target. BA45’s pristine and up-to-date property ownership records are ideal for creating highly targeted digital ad campaigns.”
A Way Back Machine copy of this SafeGraph + BA45 partnership page can be viewed here – most of the other evidence of their partnership has otherwise been purged from the SafeGraph website, except in a few places…
SafeGraph forgot to delete their mention of BA45 from this blog post about insurance data appending, and SafeGraph included a note highlighting the cost was half a penny per record, or “custom pricing for enterprise” for this BA45 integration:
SafeGraph also wrote a blog post where they highlighted what they called, the “29 best geospatial data providers for deeper analysis”… and highlighted their own partner BA45, making sure to note, “At half a cent per record, its data is very affordable as well.” :
SafeGraph briefly stopped pretending they supported anonymous data sales with their blog post, “Where to Buy Alternative Data for Cutting-Edge Insights” and mentioned BA45 immediately after their own company, noting that BA45 is “really affordable, too.”
In another blog post titled, “Geospatial Data Sources: Where to Get the Data You Need,” SafeGraph was quick to compare their own company to BA45 for property data, and made sure to note that BA45 has “a database on over 125 million properties across the US, including information regarding who owns them.”
A similar blog post from SafeGraph, “Where to Buy Geospatial Data from Reliable Providers” also highlighted BA45 and a few other location vendors.
BA45 Data included 103 Million U.S. Consumer Records in the SafeGraph Integration for “Ad Targeting”
BA45 data included 103 million records, 99% with owner last name, 100% with an appended Placekey location ID, according to their own “Property Owners Coverage” tab here.
There are many questions that BA45 and SafeGraph should answer :
- Where did the BA45 data come from that was used in the SafeGraph integration?
- Was any of the BA45 data generated from social media scraping? Facebook? Linkedin? Other?
- Was any of the BA45 data from Orb Intelligence?
- Did the BA45 data come from Dun & Bradstreet or was it shared or connected to D&B in any way?
- Did D&B know about BA45 and the SafeGraph partnership?
- Safegraph knew the BA45 “ad targeting” features because SafeGraph literally wrote it on their own partner sales pages. Why did SafeGraph tell reporters and regulators/politicians that they didn’t provide any deanonymization services, while this partner integration was active?
- Where is the BA45 data now? Has this business shut down or sold the assets?
- Was anyone else who previously worked at Yandex involved in BA45? Anyone at Yandex currently? Was the BA45 data shared or sold to anyone at Yandex or connected to Yandex in any way?
- Did SafeGraph or anyone working with or investing in SafeGraph also invest in BA45?
- How many SafeGraph clients used BA45 data?
- How many millions of records were appended through BA45 and the SafeGraph partnership?
- BA45 shut down somewhat unexpectedly, why?
Questions? Concerns? Answers to any questions above? Ping me on twitter @thezedwards