Welcome to Traction, your weekly interview series with founders who share their stories and reveal secrets behind early traction.
I am excited to introduce Dan Erickson, Founder and CEO of Viable.
I’ve been following Dan’s journey for years and it’s an incredible one. Viable has raised a seed round and is making waves among AI-first startups.
Dan was very generous with this time—he shares about their “pivot” and how they initially got off the ground.
Table of contents
— What is Viable? (1 min)
— Inklings of traction (4 mins)
— The “pivot” (3 mins)
— Getting paying customers (2 mins)
— Seeing early signs success (3 mins)
— Churn (2 mins)
— Running experiments (4 mins)
— Traction channels (2 mins)
— What he’d change (2 mins)
— Advice to founders (1 min)
📚 How to read this newsletter: Flip around, go to your desired sections, and bookmark it for later reading.
Introduction
Joshua: Hey Dan! Can you tell us who you are and what you've been working on with Viable?
Dan: I'm Dan Erickson. I'm the CEO and co-founder of Viable. Viable helps companies understand their customers or employees through qualitative analysis.
We aggregate data from a bunch of different places like surveys and App Store reviews and customer support tickets, and deliver those insights to your team to help you improve your product and your processes.
Inklings of traction
Joshua: And so when did you first start seeing signs of traction for the first iteration of your product?
Dan: In order to answer that, I gotta go into history a little bit….
So we actually started the company off with a very different target market in mind, and a slightly different value prop.
We knew that we wanted to start something, and we knew that we were going to have to find product market fit in order to do that.
So I started to research a bunch of different means of helping measure and improve product market fit, and came across this article by Rahul Vohra over at Superhuman that outlines the process that they took to understand their customers well enough to improve and measure their their product market fit.
It's a survey that you send out. You ask about four different questions in there, and you use the answers to those questions to help guide your roadmap, to help you improve product market fit.
We did some research and found that a ton of companies were attempting to use this. But not that many companies were actually sticking to it.
It’s a human time intensive process that you have to run. In fact, Superhuman spends about a dozen hours a week, just manually combing through their data.
And tagging and sorting that data. Most small companies that don't have product market fit yet don't have that kind of time to dedicate to it.
However, this was back in in late 2019, and natural language processing was just getting to the point where we could actually start to rival humans at that tagging and sorting tasks.
So we actually launched the company as Viable Fit back in January of 2020.
We built that the initial version of the product up over the next six months, got some initial traction with some some early stage companies.
We launched on Product Hunt had over 500 companies sign up to use us. But we quickly realized that they're just at that early stage.
There's not a whole lot of cash to go around. There's not really a big business that you can build off of pre-product market fit startups.
So that was a quick realization.
And luckily we found that within the first six months of the company but we also started just see usage from companies that were much larger that clearly already had product market fit.
So we quickly changed market to growth and enterprise stage companies, and actually stopped doing the product market fit process survey.
We completely retooled our product to to cater to the needs of these larger companies and build a whole aggregation system to take pre existing qualitative data and pipe it in as well.
The “pivot”
Joshua: I remember you telling me about your pivot. Can you talk about what happened?
Dan: So first off, there's a bit of a negative connotation around the word pivot. And I have a good sense for what sort of investor sentiment is and oftentimes if you talk about it in terms of a pivot, investors and potentially even customers, think of it as kind of a company reset.
They think that you hit the eject button and got rid of all the learnings that you had up until then, and that you burned all that money, and now you're starting fresh.
And in reality, that's really not what most people mean when they say that they're pivoting.
What they usually mean is that they had this initial idea, and they used it to go to market and learn a bunch of stuff from that target market.
Like we did, we learned that a lot of these companies don't have a whole lot of volume of data coming in. So there's not a whole lot that we can help them automate.
We learned that these pre-product market fit companies don't have a whole lot of cash to spend on this on this kind of thing.
And then on top of that, a lot of them don't even know sort of when to send the right and like when is the right time to send these surveys.
They don't have enough engagement from their product to actually even get the responses that they need. So there was a lot of things that we learned in that early period that showed that our initial assumptions were off.
However, we also learned from people who worked in our target market that ended up trying us anyway.
That actually ended up pulling us upmarket to a totally different market. It was that learning actually, that was the most valuable process or valuable thing to come out of that first six months of the company.
Getting paying customers
Joshua: Did you have customers before the retooling that you had brought over?
Dan: We had at the time about a dozen or so paying customers.
Joshua: Were they overall positive, negative or just neutral to the changes?
Dan: Overall, mostly neutral, which is in itself a good signal that you weren't quite onto something there yet. We actually tried to make that transition as smooth as possible.
Whenever you're changing the ground out from under your customers, you got to make sure you're making good on your promises to them.
And so we actually worked with a few of them to set up processes internally to help them run the Superhuman PMF process without our tool, and we actually offered to help set up Typeform surveys, and then pipe that into the actual the new viable product to to do the analysis.
While we pivoted away from away from that particular market and that use case, we still did our best to help support that use case for those existing customers.
Seeing success
Joshua: When did you actually know that you were onto something with with Viable?
Dan: It’s been a gradual process, and I don't I don't know if there's any specific point I can point to that says, “yes, I know when.”
However, finding your core value, I think is that is the first step. So finding the thing that customers are actually reaching out to you for, what the problem is that you're trying to solve, or rather, what they're trying to solve with your tool.
It was a culmination of a bunch of little signals that we got from a bunch of different sources that slowly built up into this picture of what I would call proto product market fit.
So, we we built up a system that allows you to sort of aggregate data from a bunch of different sources, and then our initial version of the product, or rather, our second version of the products.
The first one post-PMF was just q&a. Basically, what you could do is you would aggregate all that data, and then you could ask a question in plain English of that data set, and we would generate an answer using the data.
So for example, if you ask the question, “how can we improve our calendar feature for our customers?” Our system would then go in and find all the feedback about calendars, aggregate all that into one place, analyze it for different themes that was found in there and then actually write up a paragraph to answer the question.
So we found that initial initial excitement for this was off the charts…
We ended up having a ton of people reach out to us because they wanted to see what answers they could pull out of the system.
We quickly realized that once once you ask all the questions that you can think of to ask, you really have no more reason to use the tool.
However, there's a ton more insights and answers that's in that data set. You just don't know what to ask. There's all those unknown unknowns. And so once we figured out that was a problem, we moved to a secondary feature of the product called Reports.
Basically what that does is it shows you all the unknown unknowns.
It aggregates all that data into one place, and then finds the themes in the overall data set and exposes those themes to to our customers along with in depth and actionable analysis for each theme.
Churn
Joshua: So you were saying that there was some churn—people would use Viable and soon bail?
Dan: Yep, and that was that was key learning from just customer feedback that we were getting from our own customers.
They'd be like, “man, this is amazing.” They might pay for like one month or at a time or two months.
And then they stopped stopped asking questions because they just didn't know any more to ask and we learned that through just talking with them. I think they the best thing that you that any founder can do.
To really understand what what the market needs for them is to talk to the market. Go talk to the customers. And at first that needs to be at you know, a very intimate scale.
Running experiments
Joshua: So as you got off the ground, what were the specific growth experiments that you could walk us through, or are there any that stand out in your mind?
Dan: I think our most lucrative one that we've done is investing in our partnership with OpenAI.
So Viable is one of the first companies that got access to GPT-3, along with a handful of other models from OpenAI. And we have cultivated a really great relationship with them. To the point where we're actually featured on the GPT-3 signup page.
We have old articles written about us on the GPT-3 blog, and that is actually still the number one referrer for us.
The next best thing that we've done is basically just a bunch of content.
So we did a webinar about a little over a year ago which actually had some some pretty good responses there. It ended up requiring a lot of manual work on our side, so we haven't replicated it yet. But we're considering popping it up again.
And then we also wrote a bunch of content about applied AI, specifically within the feedback space.
But generally speaking, we're actually finding that our more technical content is actually providing more value for us right now… more than our less technical content.
So we're actually going to be doubling down on that.
Joshua: And and so when you say you're creating content, what channels are you posting this content on—is it just on your own blog or social media?
Dan: It's mostly or our own blog, and then it is blasted out to both Twitter and LinkedIn.
And then on top of that, we also do a weekly newsletter to to our entire newsletter list, which includes all of our customers, and anybody else who has expressed interest.
And that weekly newsletter goes out with not just content that we wrote, but also interesting content in the qualitative space, as well as any new feature releases or anything like that.
Traction channels
Joshua: And then what kind traction channels are exciting to you?
Dan: This will come with a little bit more investment in scale on our side but we're actually looking at one thing that that our product does that no other product does, is it generates actual written analysis. There are a lot of public datasets out there that could benefit from having this analysis done on it.
Things that I'm thinking through here are things like apps, App Store reviews, right? Imagine going to the top 100 apps on the Play Store and the App Store and then doing an actual analysis of their reviews and publicly posting that on our website.
Each one would have its own highly SEO page to talk about, what are the top compliments, the top complaints, the top record feature requests, and the top questions that their users have.
You'd have one place that you could go to find that for all of the top apps in the App Store. We can do very similar things for the top Subreddits for instance, or any other sort of public dataset.
In hindsight
Joshua: What would you have done differently kind of looking back at the last few years?
Dan: Sure. I think this is probably a common answer across most most startup leaders.
I look back and I see some decisions that I made quickly and decisively like moving away from the product market fit product and towards the enterprise.
But I also look back and see other decisions that that I just took too long to make.
So things like bringing on a sales hire. I actually waited way too long to do that. And to do that I have an engineering background and my co founder has a product design background, and we kind of meet in the middle of product management.
I knew it was a weakness of mine and was filling it in.
I think rapid decision making is the thing that I would have changed. We'd be a little bit further ahead right now.
Advice to founders
Joshua: What advice would you give to founders who might be new to this, it might be their first startup, or perhaps to someone just trying to get something off the ground?
Dan: So this ties back into the where we started the conversation actually, and about this whole concept of pivoting or following the market. My advice is your your initial idea is almost certainly a bad one.
However, you still need to implement it and you still need to get get people using it because that's how you find good ones. Building a company is not about having a vision and then building towards that vision. It is about learning what the market needs and building that.