5 lessons learned and tons of experience gained by first-generation Latvian startuper and co-founder of Reach.ly Ernests Stals.
The co-founder of one of Latvian flagship startups Infogr.am , Uldis Leiterts once remarked that he believes startups develop in generations. Uldis is one of the first-generation Latvian startupers, along with Ernests, the co-founder of Reach.ly and TechHub Riga.
This post originally appeared on LinkedIn and had this song added as soundtrack – So if you want the ultimate reading experience, play it on the background.
Failure is something awesome when it comes to business literature and stories about successful people. It is part of a great story where there is a hero who stumbles upon challenge, overcomes it and comes out as a winner. In reality, when you fail and hit that low, there is no winning part! Not yet! You have just failed and that’s it. You feel bad and it sucks. People around you might like to read about failures but, in real life, they won’t be that passionate about you. You are just a loser. Mark Suster thinks it even smells.
Below are my hands-on failures I have had at Reach.ly. In no particular order. More of a list for me to come back and check if I’m doing something wrong again.
Cap Table
Reach.ly really had two totally different iterations – Twitter mining solution for hotels and Behavioural analytics for e-commerce. I didn’t start fresh but instead kept all previous angels. At the time, it seemed to be the right decision. I personally still had a decent amount of shares. What I didn’t have was room for other investors. Not at a reasonable valuation. As it happens I got into this situation slowly. Previous idea didn’t fail straight away, I got some money from consultancy jobs and could hire back some employees. I got some funding. We just kept on moving till in one moment you start to talk with VC’s, they ask for Cap table and you realise that it sucks.
Lesson learned: New adventure should be started as one.
Technology
We started out with Big Data stack. As we were tracking on-site visitor behaviour we thought that there will be a large amount of data and we will need appropriate stack which can handle the load. I used to say “If it doesn’t scale, it fails”. We didn’t start with a simple stack to solve a business problem and test the idea, we started with a stack that solves the scaling problem. Ironically, the crucial point was the availability of resources. I got startup deal from Softlayer which gave us 1000$ worth of hosting per month for free for a year. It is “all you can eat” buffet for developers. As a result, we had a cluster of nodes running Hadoop and HBase, logging system, monitoring solution and all the dependencies in the middle. Very close to 1000$ spent per month on hosting without a single customer or working prototype for that matter. More on technology and failure here.
Lesson learned. Start small, build on it.
Pilot clients and product-market fit
In the beginning we had pilot client. E-retailer with proper size and amount of data to develop behavioural analytics and with potential to do whatever we want (in privacy policy limits) with data. As we were so much into Big Data technology, we didn’t create appropriate tracking solution and had holes in our dataset. Instead of fixing it, we decided to switch to Shopify platform because it has standardised API and app marketplace. Sounds great but it wasn’t. As we discovered later, Shopify players were too small for our solution and we couldn’t get a decent one to work with us and become a pilot. Without a pilot client, we became like minions without a master to serve.
Lesson learned. Always have someone who can give you fair feedback on your prototype. Always.
Team
First of all, there is no such thing as bad employees. There are just bad hiring decisions. I had a team which was technology centric and were not that much interested in the actual business side. As the CEO, I was the one who got them on board. They were interested in working for Reach.ly, and so it seemed to be OK. I didn’t evaluate their personalities, their goals and how they fit with company goals. As a result, there were big misunderstandings. In the end, some people left and I finally let go others.
I spent half a year hiring people who fit the company, the team, the product. The new team could iterate really, really fast. In retrospect, I should have let people go much earlier and halt operations till I get a proper team on board. It comes down to people who can work together as a team. If your team cannot ship, there are problems. Even worse, there are greatly missed deadlines and, as the CEO, I become the front-facing over-promiser.
Lesson learned: Passion [really] matters more than knowledge.
Investor communication
In startup environment, a lot of attention is on pitching. You sell yourself, your idea, your team, your vision. After the sale comes the communication part.
In the very beginning, I didn’t communicate at all. There were so many delays in development that it just didn’t matter. I wasn’t required either. But once I got our team in place I started to communicate. I followed simple structure for my monthly reports:
- Summary
- Accomplishments
- Let downs
- Goals for next period
One thing I missed was financials. Burn rate and runway. The structure above is focused on product and market. On the other hand, money is the fuel for all of this and gives a reality check for everyone.
Lesson learned. Keeping everyone on the same page is crucial in the long run.
The good parts
Thank you’s
There are some people, groups and institutions which have put trust in me and I’m thankful for that.
- Imprimatur capital – without them this wouldn’t have happened. Toby an
d the team has extensive experience with early-stage startups in Latvia and Baltics. - Andris Berzins – mentor and advisor who is now turning into VC. Thank you for pushing.
- Super secret startup Latvian CEO group – Thank you for critical feedback.
- My team – we have moved forward to new venture – Dripit.io.
These lessons are not unique. People have written before about them. I just had my own way arriving at them!
Next up
While working on Reach.ly, we noticed that e-commerce analytics process is still very labor-intensive, even where it shouldn’t be. People need to make decisions based on data, not spend hours digging through their datasets for clues. For each insight, currently you have to take these steps:
- Know what to look for. Figuring out the question can actually be even harder than finding the answer.
- Pick the right data. Miss one data point and your analysis may be flawed or even return nothing.
- Do the actual analysis.
- Hope it is done at the right time. Some insights are daily, like spikes in traffic from a particular source. Some take time to notice, like change in loyal customers.
Drawing on our intimate experience with e-commerce data, we have built Dripit.io – a solution that tracks hundreds of data points and picks up patterns and trends for e-retailers. Results are presented in a beautiful timeline, similar to Twitter and Facebook insights.
Reach out to become a Beta tester. For Shopify users it is already available here.
