Numinar - William Long
Today we have William Long, founder of Numinar Analytics, for an interview. Numinar is an AI political campaign CRM that leverages data-driven targeting and omnichannel outreach to support political campaigns of all sizes.
Brian Wei: Tell me about yourself and how you got into the tech ecosystem.
Sure! So I’m originally from Oklahoma and have been there my whole life before I got into Harvard. At Harvard I did a joint concentration in computer science and government and one of the first things I did was jump into the hackathon scene--I thought it was a really cool space. Then I got involved in the innovation lab. It was really interesting to see people who don't have years of academic experience, a PhD, or technical training build really cool and tangible products. I love that aspect of it. I really dived into that ecosystem and ended up interning at Amazon, doing machine vision for the Amazon Go Store one summer and, after junior year, I ended up spending about eight months at Palantir in D.C., working as an engineer from there. All of those experiences just taught me so much about how to build good products, as well as the hackathons in particular--I had done a ton of those. I love how you pretty much go through a mini entrepreneurial journey in each hackathon, coming up with an idea, building a team, developing the product, and pitching it in 48 hours or so. So I ended up doing ten to twelve of those during my time in college.
I understand that you founded Numinar to provide data-driven campaign software. What encouraged you to start the company in the first place?
Going into my junior to senior year, I discovered a love for humanities. As you know, I came in with a tech-heavy interest, but ended up discovering government--political philosophy in particular--and ended up doing a joint degree in computer science and government. I did a lot of political conferences on the side and did some campaigning as well over the course of 2016. That process exposed me, for the first time, to the world of political campaigns. As someone coming from Silicon Valley background, the software we were using and the way we thought about data in campaign strategy was outdated. So it just got me really curious about the state of technology in campaigns. That encouraged me to do a lot of industry research my senior year.
What does Numinar do from a product standpoint?
So we do a couple of things. Largely, we wanted to build a tech platform that would essentially use voter data and modern data science to drive a highly optimized campaign strategy, targeting campaigns at the state house and senate and local level in particular. Candidates that don't have millions of dollars in funding and a ton of expertise. We have our own models that predict things like candidate preference and turnout likelihood, based on a number of different demographic characteristics, voting history, consumer data, and more.
It acts as a form of artificial intelligence. It actually learns over the course of a campaign by leveraging the feedback that we get from voters. Whenever we knock the door, send a text or call them, we store that feedback to essentially retrain and improve the model. That's pretty important because in order for you to know how to win your race, you need to know how many votes you and your opponent have and where to find new voters. So the predictive modeling helps with all of that from a data perspective, and tells you how you need to be approaching your campaign.
We also built a whole CRM system on top of the voter file data. For every client we work with, we have data on every single registered voter in their district, like names, addresses, and other demographic information. We allow them to interact with it and improve the data we have on each voter. And, ultimately, we can conduct all the voter outreach. From our platform on your computer, you can make calls, you can send texts, email, and programmatic direct mail (postcards and letters) through our software. You can also send Facebook ads and execute things like custom audiences. And, of course, we also built applications for geographic analysis and a mobile app that you can use to actually walk out and knock doors with.
Given the backdrop of misinformation and data breaches, how does Numinar ensure data is kept safe?
Yeah, totally. I think a lot of my experience with working with data security comes from my time at Palantir. We use some of the best practices out there in terms of firewalling, database management and access control. There are many technical things you can put in place to make sure that people who aren't supposed to access the data can't access the data. In another perspective from the data privacy stance, I think politics is an interesting and unique space, because it is really core to the American life. I think there are many unique privileges that we give to politicians and candidates with regard to the data they can use. It’s important to American democracy that the public is aware of who's running for office, when elections are, and what candidates stand for. Ensuring people’s awareness and engagement is crucial to the continuation of our country. We also take a lot of measures to make sure that we take care of people’s offline data. Voter file data we have offline doesn’t communicate with online data, such as digital media presences. That's pretty important to ensure that we're not overstepping any boundaries with regard to people's data privacy.
How is the interest in Numinar during this election cycle?
There has been a huge interest in it. I think people totally get what we are providing, and how it's substantially different from what currently exists in the ecosystem. We started building in June in 2019 and we ended up working on 10 campaigns across New Jersey, Virginia, and Louisiana that year.
We picked up a pretty good number of races. We tested it there and spent the summer sending approximately a quarter million text messages, 8,000 mailers, and Facebook ads. We built ads and knocked doors with it. We got a lot of initial traction there. And then for 2020, obviously, we want to expand quite a bit more significantly. At this point we have 15 campaigns signed up all across the country. There has been a pretty wide amount of interest, especially from the political committees and state parties. Ultimately, we want to do something like 50 campaigns this year.
How do you ensure there’s enough demand for your product when elections are not happening?
That’s something that I definitely have a lot of ideas about. We’re already looking at commercial and legislative advocacy applications for our software. An elected official may use the same tools to engage with their constituents or with people who care about a particular issue or a legislative advocacy group can use it to promote ballot initiatives. We also have a vision to go abroad and support elections all around the world.
How do you usually ask for feedback from customers? Is it prompted, or do they send an email/message directly to you?
Ensuring that clients are happy is extremely important. I give them my personal cell phone. I ask them to text me or call me for any feature requests. If they have things that they want, we try to be very close to them and turn that around to build it in a matter of days. That's the one thing that I think we get a lot of positive feedback on: we're on top of every piece of feedback we receive.
What are the most challenging aspects of founding a company?
It’s a lot of work--anyone can tell you that! It is an obsession, something I hadn't realized. It bleeds into every part of your life. You're constantly thinking about your business, clients, and timeline for building out a product or new feature ideas. It’s constantly in your brain, in the middle of the night, when you're still out with friends. It’s consuming in many ways, which is something I hadn't expected before.
It's always pretty tenuous and, financially, it's pretty risky. Also, there are many people who won't get your vision. There are many people who will as well.
What are some (if any) aspects of starting a company that you wish you would have known prior to starting a company?
I think I would have encouraged myself to have more product discipline. Being more focused on nailing down the question of “what is my product?” What features are part of it? What features are out of scope? I think we really did end up building everything at the same time. It worked out and we get a lot of value out of it. However, it was a ton of work trying to build everything at once. Having a clear product thesis and sticking to it with your development timeline is important.
Separately, I didn't really recognize the importance of networks in this space particularly. I mean a lot of sales have been bound up to who you know, who you're making the effort to go out and meet and talk to. I would also recommend myself to do better at the email game: following up with people, making sure that you’re responding to everything, and building relationships. At heart, I’m a product person. My bias has always been towards designing and building a really good product that will automatically attract people. But, obviously there's a lot more to it than that.