Data has to be accurate. But what is more important – accuracy or the breadth of data? Even though it is hard to deal with changes in data, its relevance is crucial. In marketing, however, the breadth of the data is more important than its accuracy.
Auren Hoffman uses Facebook as an example. He says Facebook is a channel that uses both accuracy and the breadth of the data. It is a win-win combination for this social media platform.
In this episode of Identity Revolution, Auren Hoffman talks about the mission and goals of SafeGraph, presents the benefits of Placekey, and gives predictions of the data ecosystem. Auren and his host, Cory, discuss the expansion of companies offering different services, tools, and analysis within the data world.
Click here for a deeper dive into the episode with Auren Hoffman
- Name: Auren Hoffman
- What he does: Auren is the CEO of SafeGraph and the former co-founder and CEO of LiveRamp. He is the Chairman and Co-Founder of the Dialog Retreat. He is also a Board Member of Datavant. Auren was the first investor and served on the board of BrightRoll. Previously, Auren was Chairman of Stonebrick Group, CEO of BridgePath, and CTO of Kyber Systems. He is an angel investor in over 120 technology companies.
- Company: SafeGraph
- Key Quote: “SafeGraph just sells data. We don’t sell services or software or tools or analytics or anything like that. We just sell a collection of facts to companies.”
- Where to find Auren: LinkedIn | Twitter | Summation.net
- SafeGraph: we make the data scientist’s job easier. SafeGraph is proud of a variety of clients using their collection of data. Clients’ goals range from putting a place on the map to focusing on logistics. “We have a lot of customers around logistics, moving things around, trucking. Retailers figure out not only where they should put their stores but how they should optimize that. Even to figure out what their store hours should be weekly. These are the types of things data science teams are doing. All three major telcos AT&T, T-Mobile, Verizon, are our customers, and they use us for many different things, including things like network planning.”
- The story behind Placekey. One of the hardest things regarding data about physical places is machine data together. “So we use Placekey to do that. Starting on October 7th of 2020, we opened it; anyone can use it. It’s free, it’s open, it’s easy, and it’s a simple API where you can ping an address or a partial address or address plus a POI, and you get back a simple key, it’s a simple string, and anyone can match on a string.”
- The relevance of the data is of the utmost importance. It’s quite complex to deal with a change in data, especially in a crisis, when many businesses vanish. Data, however, must be accurate. “If you think of a pre-COVID, the death rate of a place in the US was roughly 1% a month, so it’s pretty high. Post-COVID is much higher. It’s been over 3% a month. We have so many different sources because any source might be wrong out of the gate.”
There were no companies focused on selling data to data scientists
“We started SafeGraph to focus on data for data scientists. One of the things for most data science organizations, there’s a lot of awesome companies that sell them tools and software, and services, but there really were no companies focused on selling data to data scientists. And SafeGraph specifically focuses on data about physical places, but our customers are data scientists, data science teams, or machine learning teams.”
If you want to get facts, that’s what we sell
“SafeGraph just sells data. We don’t sell services or software, or tools, or analytics, or anything like that. If you want to get facts, not fake news, that’s what we sell. And pretty much a hundred percent of our clients are data science teams or machine learning teams.”
The state of the market in the data ecosystem?
“Well, there are a lot more data buyers than there were a few years ago. And the number of organizations that can buy data today is much greater than it ever has been in the past. Now, you can have a tool like Snowflake or Databricks or some of these others, and you can be as dangerous or even more dangerous than a great engineering team was five or ten years ago.”
A monopoly within the data ecosystem. Is it possible?
“Well, there’re so many different tools that a data scientist can use. And many of which solve different things. And obviously, you have incredible companies like Snowflake or Databricks just handling the data and a lot of other companies helping you analyze it in different ways or visualize it in other ways.”
The MarTech and the data
“In marketing, the data does not have to be as true as in almost every other function. If your data is like 50% accurate in marketing, that data is gonna perform well. If your data is 50% accurate in almost any other function, you will fail.”
Since 2010, the marketing landscape has gotten more fragmented
“I think that trend probably continues across every industry, whether it’s marketing or if you just think of the number of vendors your company has today versus the number of vendors you had five years ago.”
[6:32] “We have so many different sources that we get our data from because any one source might be wrong or out of gate.”
[7:35] “The number of organizations that have the ability to buy data today is much greater than it ever has been in the past.”
[8:04] “So this proliferation of tools that sell to data science and machine learning teams have increased the number of companies that can buy data today.”
[10:56] “Part of the reason that Facebook is probably such a successful marketing channel is that they have both high breadth and they have super high accuracy.”
[11:23] “If you had to choose breadth or accuracy in marketing, I’d still think a lot of people would choose breadth over accuracy.”