500 Startups Founder Bootcamp
A talk at the 500 Startups founder bootcamp on what we learned growing PracticePanther: testing your product with real people, measuring what users do, calling new sign-ups fast and using NPS to save customers before they leave.
You’ll learn:
- How the “grandfather test” shows you where people get stuck in your website or app
- Why you should measure the steps a new user takes before they become a paying customer
- Why “always be testing” beats letting the loudest person in the room decide
- Why we call every new sign-up within four minutes
- How an NPS survey finds unhappy customers while you can still save them, and turns happy ones into reviews
Resources
- Subscribe on YouTube:
@davidmbitton
Build less, learn from competitors and early adopters
The talk opens with how to build the first version of a product. Most products already exist in some form, so study your competitors, speak to their customers and learn from the mistakes they already made. Then keep your first version as simple as possible, and bring the engineers into the planning, because a small change to a spec can make a big difference to how long it takes to build.
Early adopters can use the product for free for a limited time, as long as they know when they will start paying. Every feature should have a measurable goal, and you should track who asked for each one, so you can tell them when it ships.
Usability testing: the grandfather test
This is where I take over. When you build your own product, you know it too well to see where it is confusing. So grab someone who has never seen it, ask them to use it while thinking out loud, and just watch. You can’t answer their questions.
We called it the grandfather test, because I literally sat with my grandfather: if he could use it, anyone could. The more of these tests you run early on, the better, and if you don’t have people around you, there are sites that will pay people to record their screen and voice while they try your product.
Measure what new users actually do
Tools like Google Analytics and Mixpanel take about a day to set up and show you the path from someone landing on your site to becoming a paying customer. For PracticePanther, we found that people who created their first contact and case, then a time entry, then an invoice, were the ones who became customers. So everything in the product pushed new users through those steps, and the more people went through that funnel, the higher our conversion rate.
Always be testing
You know “always be closing” from the movies. Our version was ABT: always be testing. Instead of arguing about which button, landing page or feature is better, test both and let the numbers, or your customers, decide. Don’t assume you’re always right, and don’t let the loudest person win.
We also learned that our plan to automate the whole sign-up never worked on its own. Doing the demos ourselves told us what people were looking for and which problems they needed us to solve.
Tools to watch and talk to your users
Mixpanel and KISSmetrics show you the numbers, and screen-recording tools let you watch what visitors click on. For talking to customers we love Intercom: live chat, a chatbot and support tickets in one place. If a visitor is on the site, they get a live chat. If not, it sends them an email. You can also send an automatic message after a few seconds on the website, or when someone visits the pricing page.
Call new sign-ups within four minutes
We learned from Mixpanel that people who sign up are testing a bunch of programs, and within about four minutes they’re gone. So we make a point of calling every new sign-up within the first four minutes. Get them on the phone while they’re still in your product, and sometimes you can do the demo on the spot.
NPS: find unhappy customers before they leave
After a customer’s first 30 days, a bar at the bottom of the product asks how likely they are to recommend it to a friend. Nines and tens are your raving fans, sevens and eights are neutral, and anyone lower is a detractor who is probably on their way out.
Once a customer tells you they want to cancel, they have already switched off. A bad score comes before that decision, so it is your chance to be the superhero. Every detractor triggered a Slack notification, and someone from customer service called them right away. That call is the wow factor.
Automate it: turn happy customers into reviews
We automate as much as we can. Every 30 days the survey appears by itself. When a promoter gives us a nine or ten and a good comment, an email automatically thanks them and asks them to post that same comment on a review site like Trustpilot. That is how we get more reviews without anyone lifting a finger.
Questions: usage, showing value and too many tools
In the questions, we talk about when to send the survey, asking customers who leave what would make them stay, and testing different pricing plans. One idea is to show customers the value you bring right on their screen, like how many hours the product saved them this month, or a savings calculator on your pricing page and in your demos. We also cover how to avoid using too many tools: if one product gives you value and you love it, use it, but don’t make your team log in to twelve different programs.
The tools we use to measure everything
To finish: promoters get an automatic email asking for a review, and detractors get a message asking what we can do better, although in the early days you should just pick up the phone. Heatmap tools show you what visitors click and scroll through on your site, and FullStory records a customer’s screen so you already know the issue when they ask for help. Only automate what saves you time today.
Everything feeds into Salesforce: billing, marketing and product data. It is built for sales teams, but it is easy to customise, and it lets us see which marketing sources make the most money, which salespeople have the best demo conversion rates, and what the customers who churn have in common.



