When building Textio, I knew that we could only be successful as a machine learning software if we designed a business model that optimized both usage (to get more data) and customer value. This core belief was fundamental to everything we did, from how we approached the product to when and how we raised capital. It was not always easy for others to understand our vision, particularly in the beginning before we had enough data to demonstrate concrete customer value. We had a lot to prove, and I am incredibly proud that we stuck to our vision, found partners who believed in us, and persevered through all of the ups and downs.
Below, I've outlined a few of my learnings that might be useful for other founders who are in the early stages of building out machine learning value for their customers.
Stay close to the product through product-market fit.
During our very early days and throughout our seed-stage funding round, it was important for me to personally be involved in all of the customer discovery in order to listen to the market signals. For the first year, we deliberately remained an all-technical team; I built the early data models, and my Co-Founder & CTO built our initial back end. Even as we made our first business hire and continued growing the team, I led our enterprise customer conversations and the feedback loop to product until we found product-market fit.
Enter a market where you are uniquely positioned to win.
Early in my career, I recognized the tremendously harmful impact of conscious and unconscious bias in the workplace, especially in the people space. Even before we launched Textio, I wrote a lot about the problem and its many implications. The work I did before Textio helped give us credibility in the HR space and grew my network of relevant practitioners. After Textio was founded, this network of practitioners became our first customers. I was able to invite them to test out the product and provide early feedback, even when the product was just a prototype.
Build a premium user experience.
We built Textio as a learning loop. Simply put, the more the product gets used, the better it works. We realized that for this model to truly work, we had to build a product that people enjoyed using consistently. We didn’t want to build a cheap tool; we wanted to build something that provided substantial value for the enterprise.
Great enterprise products have a premium look and feel. From the beginning, we invested heavily in UI design. We believe this investment more than paid off; even when the product was still free, people treated it as a premium offering. Consistent usage enabled the platform to continuously improve.
In the beginning, it’s worth giving something away for free in order to get feedback, even on premium products.
At Textio, we found that it was worthwhile to offer a beta product for free in return for early user feedback. This strategy allowed us to build up the data we needed in order to get the system to work and allowed us to implement our models within enterprise systems before the platform was ready for paid, commercialized use.
Everyone won in this scenario. We received critical product feedback, while these early beta customers felt like insiders and were able to use the product for free. We started charging them much later, once the product was mature—and valuable—enough to justify it.
Target large enterprise customers to quickly scale usage and data.
Our go-to-market strategy oriented well with large enterprises. For any learning loop product, the more people you have using it, the better it will work for everyone. Enterprises were a natural fit for us; they write enough documents to really need Textio, and they have enough people to make Textio learn quickly. This means that Textio works better and better for them the more they use it.
However, this was not an overnight success—it obviously takes longer to sell into larger enterprises than into smaller businesses. But once we made it into organizations like Cisco, Johnson & Johnson, and American Express, we were able to build a learning loop powerful enough to deliver meaningful customer value.
Design pricing to encourage usage.
Our product is only as good as our data. We recognized early on that charging per-user would make our pricing structure the enemy of our product adoption. Per-seat pricing would only discourage users and limit data creation. Instead, we found that flat-rate pricing fosters the necessary network effect since each additional person makes the system smarter.
Leverage customer testimonials when fundraising pre-revenue.
Early in our fundraising process, Textio did not yet have any paying customers. To get in front of potential investor concerns, we solicited testimonials from our early adopters attesting to the value of the product and stating that they would eventually be willing to pay. We attribute our success to our early customers and investors who really believed in us.
Some customers provide an outsize impact on your product.
Textio is an augmented writing platform, so our users are writers. We’ve found that the best writers are constantly pushing the envelope, seeking out new ways to differentiate themselves and developing new terminology. For similar ML platforms, I would recommend creating a system that can identify innovative users and build algorithms around the strongest usage patterns. This will ensure your network continually learns from the highest integrity data within each organization. After all, the whole goal is to help teams access their full potential.
Please feel free to reach me at @KieranSnyder (my DM’s are open). I would love to hear about your experiences or answer any questions. Thank you.
Enjoying this article?
Sign up to gain access to our thought leadership and have future articles delivered directly to your email.