In December last year, I had finished a proof concept for a chrome extension that determines people’s DISC personality type based on people’s LinkedIn profiles. The plan was to spend the following two months (January and February) on user discovery - understanding what my potential customers really need. In this post I’ll talk about how that went, what progress I made since then, and what’s next.

🔍User discovery for a personality analysis app

My goal in user discovery was to talk to 100 people by the end of February. I knew from the start that this was unrealistic. But this is a number’s game and it’s important to aim for something ambitious to at least get to something workable.

My hypothesis, based on how I’ve seen my then competitors  (Humanlinker, Crystal, Humantic) market themselves, was that sales people could get better results (i.e. conversion) from personality insights. For example, knowing their client’s probable communication style could help them adapt their own style.

I ended up reaching out to about 150 people, which resulted in about 15 useful conversations. My initial target were sales org leaders (e.g. Heads of Sales, Chief Revenue Officers). Luckily, I had some conversations with some smart people along the way that also pointed me towards other roles like Business Developers and Growth.

From people in Sales roles, I learned that personality types are cool and all, “but, like, so what?”. It’s a magic trick, but it’s too far from being actionable. There was more enthusiasm for the possibility to write messages tailored to user’s personality, but that wasn’t enough either. The key insight was that there are multiple types of insights (such as personal career track, personal KPIs & pain points) that go into writing a good message

From people in Growth & Business Development roles (one of their jobs is sending outbound campaigns at scale) I learned that one of their biggest struggles is writing the right hook (e.g. the first sentence that explains why they are reaching out.).

Another thing I learned, is that most people don’t trust AI writing their messages. The point of using AI is to make something automated sound like something that’s not automated. But even with GPT4, that’s hard. Could I find a way to make AI sound normal?

Actually, (lowkey brag incoming) this is how I got my first revenue - I was able to generate about 600 personalized hooks for a Growth client (ahem, my first client).

✨TypeCharm is live on the Chrome store

Iterating research and message generation prompting for my client helped me get to pretty good results. I was very eager to put these results to work.

So in the past three months, I’ve added the ability to generate messages to the chrome extension (now called TypeCharm). To do this, I’ve broadened the type of research it does the long with the personality type stuff (e.g. determine the prospect’s likely KPIs & pain points and figures out good talking points for an initial message).

You can try the extension here (please leave a rating and a review if you try it), and check out some stylized screenshots below.

⏩Next steps for TypeCharm

One of my struggles so far was to balance product development and customer research. Conventional startup wisdom says to first sell then build. In my experience so far, customers wanted to try a product first, before they bought anything. So now that TypeCharm can be used by anyone, I’m back to customer development and discovery.

For the next three months, I’ll be doing cold outreach (using TypeCharm of course) and posting on LinkedIn to generate some inbound as well. On top of this, results for my client’s campaign should be coming in the next month. If all goes well, that should bring in more orders 🤞.

🗿Personal update

I’ve been training for my second marathon which will happen on June 2nd, in Caen, France. Follow me live here.