Overcoming the Hurdles Facing A Silicon Valley Polling Startup

When you’re from Silicon Valley and you’re pitching business, you need to convince leaders of campaigns and causes that you understand politics.

When you’re selling something for $5,000 and they’re used to paying $40,000, you need to convince them you’re providing a high-quality product.

When you’re doing it online and they’re used to doing it on the phone, you need to convince them your way is more accurate.

This summer, we founded Change Research as a public-benefit corporation. We have a new approach to public-opinion polling, we’re charging a lot less, and we’re opening up polling to a lot more candidates, and that means we have to overcome a lot of skepticism.

Here in Silicon Valley, early adopters are willing to try any cool new technology. But campaign operatives don’t want the new thing. In fact, when we launched, a few cynics told us that campaign managers would only stick with what they knew. But we’ve been pleasantly surprised. 

After talking with dozens and dozens of pollsters, candidates, consultants, and others, here are a few of the lessons we’ve learned: 

A little humility goes a long way

When you’re new to a field, learn from the people who have been there before. Most public opinion polling is conducted by experts with many years of political experience. We know data science and we know online behaviors, but we’ve gladly learned from the deep experience of established pollsters. 

Show ‘em what you got

It’s worth investing early to show candidates what they want to see. We didn’t get into this business to build a better horse race predictor. But we realized early on that elections are the ultimate proof point for pollsters: accurately predicting outcomes counts for a lot.

In the past seven months, we’ve spent many hours and dollars running public polls that no one was paying us for in House and Senate special elections and in state races in New Jersey and Virginia. And those dollars and hours have paid off, as we can now demonstrate results.

Find bold and motivated challengers

The incumbent with millions of dollars in the bank is unlikely to bet on a new approach. Down-ballot candidates may be willing to experiment, but most candidates for state legislature, city council, and even U.S. House have less money to burn.

In New Jersey and Virginia, we focused much of our effort on strong, motivated underdogs, making the case to campaign managers, mail specialists, and challengers in New Jersey and Virginia to recruit a handful of crucial case studies. Lee Carter, considered a huge underdog in Virginia’s 50th House of Delegates district, and Judge Carolyn Nichols, aiming to become the first female African-American Democrat to win statewide in Pennsylvania, became clients. Both gambled on a new approach, and both won.

Deliver lots of stories

Publishing interesting insights is a great way to get noticed. We started publishing the results of our own national polls on Medium before we were incorporated and aggressively pitched polling media like FiveThirtyEight, The Washington Post and data journalists with our results. 

We pushed out a “crazy” poll pegging Virginia’s Republican gubernatorial primary as neck and neck and were described as a “firm out of Silicon Valley nobody has ever heard from”, but our prediction was spot on.

Feed the media hunger

When it comes to media, speed helps a lot. With so much hanging on the topsy-turvy U.S. Senate race in Alabama, any and all details about the race are newsworthy to media and interesting to political pros at large. But no one cares about yesterday’s news. You have to be fast.

In Silicon Valley, there’s a saying that a new product has to be 10 times better than the old one to succeed. While we seek to learn from established leaders, we started this company because we felt we could create a product that is far more effective at measuring public opinion.

We are in the early stages of building a massive database of survey questions that allow us to understand which ways of asking things are most effective. We want to precisely measure the effectiveness of new advertising techniques. We want to use machine learning, not simple weighting, to see how campaigns are doing.  

But we recognize that those are more attractive to nerds in Silicon Valley (and maybe a few working in campaigns) than your average campaign manager. So we aim to deliver in the short term, while building for the long term. Campaigns can be all about the now, but innovating for tomorrow, takes courage and persistence.

Mike Greenfield and Pat Reilly are the co-founders, Change Research, which provides polling to forward-thinking candidates and causes.