Powering Education Reform Through Data Science
An interview with Aaron Feuer, Founder & CEO at Panorama Education
- Tell us about Panorama as a product. How was it built and how is it maintained? How has it changed over time?
“Panorama’s product today allows schools to collect and analyze data on key topics including school climate and safety, growth mindset and SEL, and communication and leadership.
The first version of the product was some code I had written for fun to experiment with analytics in PHP. Any technical person reading this would know you should never build analytics with PHP. We were on MongoDB for a while, and I wanted to learn Python at one point so we had some components in Python. About 2 years ago we fully committed to a professional Ruby on Rails stack. There is a certain immaturity to being a software engineer in college — you want to learn everything, and it was good for us to move past that.
The Panorama Platform
While the first version of the product at most startups is written with the goal of getting the first in the customer’s hands quickly, there comes a point when you need to focus on building a sustainable product and a rigorous engineering team. How do you figure this out? Two things. The first is the moment you realize you have a lot of users in the product and it is actually worth investing in a better technology stack. When we realized that, we started making decisions for years into the future, rather than months. After a certain point, you get burned enough by bad habits, and fixing broken deploys. At some point, every good engineer starts to automate their practices, and being a good engineer means reaching a point where you buckle up everything.”
2. EdTech has some unique aspects such as long sales cycles, need for delayed updates over winter or summer breaks, etc. How did you tackle these from a product standpoint?
“It is a common misconception that education is meaningfully harder than any other industry. It is certainly not easy but it seems, at least to me, much harder to get a million students on a web app than to reach a million students through their schools. If you expect those challenges when you go in, you create them for yourself. We ignored everyone’s advice about going into the education market, which was really helpful for us. If you ignore the challenges often they are not true. Instead of building a product for challenges we had heard of, we focused on building a product that is is going to make a difference at every student in the country in a matter of years, not decades. We see companies building products for a user base that doesn’t exist. For example, you cannot build an iPad app to reach everyone in the country. We wanted to build a product that improves, grows, and betters students, for everyone.”
3. What advice do you have for someone who is building a data-driven product?
“First, you have to build it with security in mind from day one. No matter how many shortcuts you want to take, security always matters. It’s one piece you never skimp on. The other piece is that most data products that I’ve seen overbuild their backend from the beginning. Our backend right now is less powerful than it was 18 months ago, in some ways. The funny thing is we are going to be adding back in features that we had 18 months ago. If your product does too many things that no one cares about, it’s hard to maintain. As a data engineer, you often want to build out every feature you think users want, but every added complexity increases the chance your product will never reach those users. Make sure you have a need for it before you start.”
4. As you’ve continued to build out the product, how do you choose what to include and what not to include? (ie how do you manage growth?)
“I will admit, we have not gotten our product process right yet. Our engineering process is stable, our product process is not. Originally, we focused too much on the features our users asked for, rather than the underlying problems they wanted to solve. And now we focused on the 10x feature versus 10% feature. Instead of building a thousand bells and whistles, we focus primarily on high impact projects, those we want to have a year from now as opposed to all the small things we wish our product did tomorrow.
I spend most of my time talking to clients. When I ask them what keeps them up at night, I reframe back to them a menu of choices, allowing them to help me pick which one is most important. If you ask a user if they want you to build a certain feature they will almost alway say yes. You need to find a way to discern which are the most important.”
5. What’s next on the product roadmap?
“Right now were are in the journey most data companies encounter make where first we wanted to get good data, but now need to figure out how we make it most useful. Think of us as a Mixpanel for schools. We have basic reporting, but basic reporting and analytics are very different. Most products that say they’re analytics platforms but are really just reporting. Our challenge is pushing into the analytics space and figure out how we actually make more sense of this data beyond simply showing it back to users.”
The Panorama Team, Spring 2015
6. What is something unique about Panorama (either the company or the team) that not many people know?
“About a year and a half ago, there were 6–7 of us on the team, and we had a company-wide retreat to determine company values. We thought most companies had stupid values. Who is going to argue against integrity as a company value? They are so obvious and they didn’t actually say anything. So we chose 10 values that were non-obvious and potentially controversial in that a reasonable person might build a company that doesn’t have that particular value. One example is that we try to embed humility into everything we do. From the interview process and beyond, everyone in the company needs to approach the world in that way. If you come by the office you will see the rest. We printed them out and hung them on the wall.”
Natalie Bartlett is the Community Lead at Rough Draft Ventures. You can find this post, as well as additional content on the Rough Draft Ventures' blog. You can also follow Natalie (@Np_bartlett) on Twitter by clicking here & Rough Draft Ventures (@roughdraftvc) by clicking here.