Something can always be learned from a little experiment. And sometimes, that something is huge. 

By Laura Behrens Wu (Founder & CEO, Shippo)

This is part two of a series from one of our How To Conference speakers. Laura Behrens Wu spoke on our panel How to Create Value with the Right Business Model. Read part one

Doubling-down on customer support early on galvanized Shippo’s learning in a tremendous fashion. We chose to experiment with live chat as an initial way of getting to know our users — small business owners — and better understanding their use of our product. When we first started live-chatting with visitors to our site, we thought we’d merely be answering questions in an attempt to help convert them into paying customers. As it turns out, our experiences with live chat became a discovery process that informed development of our brand, product-market fit and initial customer base.

Early on at Shippo, we started using Olark, adding a simple live chat window on our website. This allowed customers to ask us questions while in their Shippo accounts, saving them the trouble of emailing or calling us to troubleshoot a simple issue or navigate the new service. Customers asked simple questions about our pricing, or sometimes logistical questions about our API or formatting labels for printing. But whatever the crux of the question, live chat enabled us to catch people in the throes of decision: We were able to talk with customers at whatever instant they decided they needed us most. This moment was the one during which their pain point was at the forefront of their minds. We were able to successfully target that pain point precisely and directly because people were perfectly primed for us to convince them to give Shippo a shot. This led to lots of initial sign-ups — gold for seed stage startups!

Since live chat was often the essential last step in convincing customers to sign up as paying customers, this process became good sales practice for us. Our goal was to determine the customer’s true problem and determine if and how we could truly help them. We aimed to be authentic, open and informative in our demeanors rather than pushy. This practice allowed us to lend a voice to our product and a face to our company. It also made onboarding a more personal process rather than a mechanical one. We found that faster, meaningful replies over live-chat quickly led to a sizeable number of happy early adopters.

At the same time, answering customer questions helped us understand what our customers were looking for, what they liked and disliked about Shippo, and what was unclear to them about our service. I was personally available on Olark chat much of the time and had a fantastic time interacting with our customers in such a deliberate and personal way. In fact, our entire team took turns jumping on live chat; every team member was given a time slot during which they were responsible for fielding customer questions. Our experiences helped directly inform the development of our first set of FAQs and our first formal customer onboarding procedures.

Looking at our website now, you may notice that we no longer have live chat enabled. After our last redesign and the hiring of our first full-time customer support ninja, we decided to turn Olark off. The months of live chat had provided us with countless helpful lessons and a faster path to product-market fit, but the scalability of our support processes required a new model. Currently, our customer support happens primarily over Zendesk and email as well as phone and social media. Operationalizing a system that improved support for all of our customers, while maintaining personalization, has been a better use of resources.

Improving those resources is an important and measurable outcome from our live chat experiment. Far more important is the fact that all of our early team members spent time interacting directly with our customers, while Shippo was still quite young. This helped us root our company in customer service as we grew. As I recently wrote, the right business model won’t land in your lap. Instead it comes with continuous iteration and good listening. The best way to get your early business model right is to let your customers inform your practices.

This post originally appeared on the Shippo blog