The End of the Data Broker Model and the Rise of the Data Marketplace
Most commodity markets have evolved from traditional broker models to more transparent, efficient transactions with lower fees and higher impact for both buyers and sellers. Travel agents, stock brokers, and even real estate agents (to name just a few) have been either completely replaced or greatly disrupted by digital marketplaces. Gone are the days of paying brokers high fixed or percentage fees for the straightforward tasks of connecting buyers and sellers to transact on largely standardized products or commodities.
Take for example the changes made in financial services: the rapid growth of passively managed, low fee Exchange Traded Funds (ETFs) vs. the more traditional actively managed mutual fund, with fund managers acting as intermediaries in the construction and management of the fund’s underlying assets.
I have seen this transition in digital advertising as well – the rise and decline of digital advertising networks. As with other ad markets, digital networks delivered scale, performance optimization, and rudimentary quality controls to digital ad buyers with non-transparent percentage fees on advertising spend. As ad formats commoditized, ad buyers found not just scale, but dramatically more transparency, quality control, and lower, more transparent fees through ad exchanges. The digital networks that captured the lion’s share of billions in digital ad spend outside of Google and Facebook in 2015 are essentially gone five years later.
This same phenomenon is now underway for buyers and sellers of data assets, and not just for martech and advertising use cases. According to the latest report by The Eckerson Group, The Rise of the Data Exchanges:
“Data exchanges are emerging as a key component of the data economy. By connecting data suppliers and consumers through a seamless experience, data exchanges have created a modern-day goldrush to help modern, digital organizations address their insatiable hunger for data. And, by eliminating the friction from finding, acquiring, and integrating data, these exchanges make it easy for any company to monetize its data assets and create new revenue streams.”
The rise of these data exchanges or marketplaces, and the decline of the broker model for data will benefit not just buyers and sellers, but more importantly, consumers that own the data they create through their digital interactions.
Before outlining the specific problems with the broker model, it’s important to ask what the key drivers are for buyers and sellers; how do most data transactions work today? Most transactions are managed by an intermediary broker through a process like this:
1) Broker asks the buyer what they are trying to achieve, e.g. better personalize an app experience, estimate the future value of a customer, understand store location visitation patterns, improve the relevance of an ad
2) Broker aggregates data assets (e.g. location data, mobile device IDs related to app installs, behavioral data derived from browsing behaviors of mobile app IDs) from a few sellers
3) Broker delivers that aggregated snapshot back to the buyer mapped to a list of emails, mobile IDs, geo locations, or similar
This method of transacting has several challenges for buyers and sellers, each of which are addressable if you fast forward beyond the data broker model.
Problem #1: Speed. Almost all the problems buyers are looking to solve are ongoing problems. Data assembled as a snapshot is stale the day after it’s delivered, and erodes quickly from then on. If you want to continue to personalize your app, you need current data all the time.
Solution: A robust data marketplace automatically manages updates to underlying data so the buyer has effectively “always on” access to the latest and greatest. There should be no need to rerun queries, normalization, etc. every time you refresh. The additional upside is that as soon as the time is cut down between the buying and selling of data there no longer is the opportunity cost that is associated with time. That opens up the ability to test, play, experiment, and change strategies as the business warrants.
Problem #2: Quality. Accuracy, completeness, recency, is generally opaque to the buyer. The buyer does not know what she’s getting even after she gets it. If the downstream results of the exercise are poor due to any of those quality issues she’ll have a hard time figuring that out, perhaps assuming the exercise had limited potential ROI based on the results.
Solution: An effective data marketplace or exchange scores the quality of underlying data assets to make critical quality attributes like accuracy and recency transparent to the buyer along with the identity of the sellers.
Problem #3: Discovery. In the broker model, buyers rely on the intermediary to find the right data on their behalf, without understanding the comprehensive catalog of what data sellers and solutions may be available. What may be “good enough” from the broker’s perspective this translates to missed opportunities for both buyers and sellers.
Solution: A true data marketplace enables buyers to easily and transparently search and browse the universe of data products that may help them improve results for their use case.
Problem #4: Non-transparent economics. Data brokers may charge arbitrary and/or high fees that obscure the underlying economics to both buyer and seller. This limits the potential value for both buyer and seller, which limits the overall potential for the marketplace.
Solution: An efficient data marketplace charges reasonable, transparent fees, and only for the data the buyer truly wants to buy. The buyer chooses the most promising data for their current application, for economics that are transparent and repeatable.
Problem #5: Compliance. Brokers have limited tools and incentives to help sellers and buyers understand source data compliance with applicable privacy and other regulatory issues, particularly as policies and regulations evolve over time. This is a huge problem for consumers. Any given data source may have unknown compliance, or worse yet, be out of compliance with critical consumer privacy protections. Consumers lose transparency and control over their data with no way to get it back.
Solution: a privacy-centric data marketplace not only promotes consumer choice and transparency with features like central opt outs, it enables all participants to track policies as they change over time, and automatically flags any data that has an unknown state of compliance.
Beyond the critical improvements needed for consumer privacy protection, the downstream effect of the problems together is the negative impact on buyer ROI. One-off transactions with low quality, potentially non-compliant data on top of unclear underlying economics means most broker managed transactions are unprofitable. Consumers do not get the more personalized app experience, the more relevant content or ad that the data buyer was trying to achieve. How much more is the buyer going to invest in these transactions?
By addressing these problems, a data marketplace or exchange that replaces brokers delivers better ROI for buyers, more demand for sellers, and better outcomes for consumers along with the critical improvements to transparency and choice mentioned above. As a board member for a company involved in creating data marketplaces, I’ve had the first-hand opportunity to see how an effective data marketplace can function and deliver these benefits for all participants in the marketplace.
Data brokers are going the way of travel agents. They’ll be around for a few more years, but if you’re looking for sustained results you’re much better off using a marketplace that delivers quality data quickly, transparently, and with verifiable protection for consumers’ privacy.
About the author: Ryan Christensen is a tech executive with 25 years of experience in enterprise, advertising, and marketplace technology businesses. Ryan was most recently SVP of Platform Product at Xandr (AT&T) after serving as the COO of AppNexus, which was acquired by AT&T in 2018. He currently sits on the Board of Narrative.io, a software company simplifying the acquisition and distribution of data. He can be reached on Twitter via @r_y_a_n_c and at https://www.linkedin.com/in/ryanechristensen/.
July 23, 2021
- Observable Introduces Data Visualization Stack for the Enterprise
- Collibra Named a Leader in Data Governance Solutions by Independent Research Firm
- Mindtech Raises $3.25M to Accelerate Growth of Synthetic Data Training Platform for AI Vision Systems
July 22, 2021
- LigaData Now Certified on Cloudera Data Platform
- Adobe Launches Adobe Analytics for Higher Education to Advance Digital Literacy
- Anaconda Releases 2021 State of Data Science Survey Results
- Alation Supports Next Generation of Data Enthusiasts, Provides Free Software and Training
- NASA Expands Access to Planet Data to All US Federal Civilian Agencies
- Google Cloud Announces Healthcare Data Engine to Enable Interoperability in Healthcare
July 21, 2021
- Deloitte, US Chamber of Commerce Report Highlights How Public Policy Can Enable Trustworthy AI
- WHO Issues First Global Report on AI in Health
- New Data Science Platform Speeds Up Python Queries
- Teradata Joins TM Forum to Support the Cloud Journeys of Global Communications Services Providers
- Confluent Named Google Cloud Technology Partner of the Year for Third Year in a Row
- Lucata Raises $11.9M Series B to Introduce Next-Gen Computing Platform
- Securonix Announces ‘Bring Your Own Snowflake’ Program to Power Security Data Lake for Snowflake Customers
- Rensselaer Team Aims to Pave Way for Robust AI in Medical Imaging
- Dremio Launches SQL Lakehouse Service to Accelerate BI and Analytics
- Spectra Logic, StorMagic Announce Active Archive Repository for Video Surveillance
- Quobyte Releases Hadoop Native Driver to Unlock the Power of Analytics, ML, Streaming
- Big Data File Formats Demystified
- Who’s Winning In the $17B AIOps and Observability Market
- Tuplex Gives Python UDFs a Performance Boost
- What’s the Difference Between AI, ML, Deep Learning, and Active Learning?
- Presto the Future of Open Data Analytics, Foundation Says
- The Data Mesh Emerges In Pursuit of Data Harmony
- Why Data Scientists and ML Engineers Shouldn’t Worry About the Rise of AutoML
- What’s Holding Us Back Now? ‘It’s the Data, Stupid’
- Achieving Data Literacy: Businesses Must First Learn New ABCs
- Composite AI: What Is It, and Why You Need It
- More Features…
- Hiring, Pay for Data Science and Analytics Pros Picks Up Steam
- Starburst Backs Data Mesh Architecture
- Confluent Raises More Than $800M in IPO
- Data Prep Still Dominates Data Scientists’ Time, Survey Finds
- Off the Couch: Database Maker Seeks $160 Million In IPO
- Let It Go: The Financial Benefits of Data Deletion
- Global DataSphere to Hit 175 Zettabytes by 2025, IDC Says
- Confluent Files to Go Public. Who Could Be Next?
- Databricks Edges Closer to IPO with $1B Round
- Teradata: We’ve Outsourced Some Hardware Support for Years
- More News In Brief…
- Splunk Launches New Security Cloud
- Red Hat Expands Workload Possibilities Across Hybrid Cloud with Latest Version of OpenShift
- JetBrains Announces Datalore Enterprise, Enabling Data Science Teams to Collaborate On-Prem
- Alluxio v2.6 Release Brings Performance, Ease of Use Improvements to AI/ML Workloads
- DDN Selected by Bytesnet to Provide ‘Pay-Per-Use’ Storage for Data-Intensive Organizations
- Yugabyte and Hasura Integration Provides Cloud Elasticity, Frictionless Application Development
- Vertica Announces Vertica 11, Delivering on Vision of Unified Analytics
- TigerGraph Expands Partner Ecosystem to Bring the Power of Graph to More Businesses
- MariaDB Collaborates with AWS to Deliver SkySQL on AWS
- Collibra Announces 24 Gold and Silver Partners for 2021
- More This Just In…
Sponsored Partner Content
August 25 @ 12:00 pm - 5:00 pm
November 29 - December 3
December 6 - December 10San Diego CA United States