Meet the owner: Wolfram Hempel

Hi Wolfram, give us a quick introduction of yourself!

Sure :-) I started as a dev in multiplayer gaming, but quickly moved on to financial markets. I've built trading platforms, automated order routing, and stuff like this for investment banks across the world. I mostly worked in London, but also spent a lot of time on-site in Sydney, Johannesburg, Copenhagen, Zurich, and a bunch of other places.

Eventually, all this moving around got a bit tiresome, so I convinced two colleagues to come with me to Berlin and start a company. It was called "deepstream" - and it sold a realtime data server to banks, gaming companies, auction houses, and similar customers.

We've raised a million in VC funding for it and got quite far - but failed to make it a financial success. So ultimately, we had to shut it down.

This left me in a tricky situation. I had sunk all the money I saved up from my previous jobs into this company - and I needed to generate income fast. But at the same time, I've tasted the sweet freedom of entrepreneurship and had no desire to return to employment.

So I started building Arcentry...

What is Arcentry?

Arcentry is a tool to create cloud architecture diagrams. It comes with a wide range of AWS, Azure, GCP, and Open Source components. It also has interactive elements like charts or metrics that you can place directly into the diagram. Now, this in itself isn't unique. What makes Arcentry special are three things:

• The diagrams look neat. I've built a 3D engine that renders an isometric style of diagram that looks good enough to plug into any board room presentation.
• The diagrams are programmable. Arcentry has an API that lets you change any aspect of a diagram in realtime. That makes it possible to show monitoring metrics, elastic scaling, or errors within their context.
• Arcentry can create diagrams automatically. You feed it a simple list of all your components, networks, security groups, etc., and Arcentry creates decent-looking diagrams that sometimes even make sense.
  
  

What was your growth strategy to set up Arcentry for success?

Content Marketing! I wrote a lot of articles about cloud concepts: serverless, edge computing, service meshes, and so on - all illustrated with Arcentry diagrams. I've published them on Arcentry's blog and shared them on Reddit and, most importantly, Hackernews, where about half of everything I wrote reached the front page after a few attempts. This brought in a large stream of relevant visitors that tried the product and ultimately upgraded to the paid tier.

Arcentry also was inherently viral. The diagrams were used in presentations and blog posts and attracted attention. To accelerate this, I've built an embeddable 3D viewer that lets you embed interactive diagrams into a blog post - of course, with a backlink to Arcentry's website.

On top of that, I used my network. I convinced e.g. AWS to list Arcentry as an official tool for AWS diagrams. Fortunately for me, the list is alphabetical, so Arcentry got the first spot.

How did you find your first customers and convince them to sign up?

Also Hackernews. I wanted the product itself to make the homepage - but that proved really tricky to do organically. I tried different titles and posted repeatedly at different times, but it struggled to gather the momentum necessary to break out of the crazy torrent of HN submissions.

Eventually, I got frustrated. I still had the email of one of the Y-combinator moderators from an earlier exchange and very politely asked if there's any chance he could give it a nudge. He agreed (thanks again, dude!), and it made the frontpage.

This sent about 60k of interested visitors to my page, which was more than enough to get the ball rolling.

You had an acquisition offer on the table 8 months after launch! How did that happen?

Not just one, actually. Arcentry's ability to create diagrams automatically and wire them up to cloud monitoring tools clearly hit a nerve and three large tech companies reached out early on.

Having said that, it's always hard to judge how serious these sorts of interests are. With deepstream, we also had acquisition talks and flew back and forth to San Francisco to pursue them - only to see them vanish into thin air over time.

My advice for any founder engaging in acquisition talks is to push for a letter of intent early on. This is still not a binding contract, but at least it keeps the "we're just exploring options phase" short and puts some concrete numbers on the table early on.

What made you consider a personal liquidity event already at this stage and which options did you contemplate?

Short answer: I was broke. Badly broke. All my savings were sunk into my previous company, which didn't work out.

It got so bad that when I got an acquisition offer for Arcentry from one of the FAANG companies, they flew me to the states to negotiate and interview me. Over there, I hung around the outside of Starbucks windows for free Wifi since I could neither afford the roaming costs nor the coffee inside.

The hotel I stayed in was paid for, but I had to put down my personal credit card for potential room service charges. When I returned, the money on the card was still blocked, and that was the last bit of cash that I could access. I was too proud to borrow anything, so I went without food for a few days.

The challenge with all this was that I had to appear successful to prevail in the negotiations. Not the easiest thing to do when you meet at a restaurant and sit there terrified, knowing that you couldn't pick up the tab if you were asked to.

You ended up selling Arcentry in a licensing deal for a seven-digit amount half a year later. Somewhere in between that and the original offer, you also managed to attract an offer from a FAANG company. What was your decision-making process leading up to the final deal?

I want to be a founder for one main reason: Freedom!

Don't get me wrong; there's plenty of other stuff I like as well: I deeply enjoy building products that people use and love. The money can be great if things work out. Plus: I love the lifestyle and naturally thrive in the chaotic, ever-changing environment that comes with the job. But first and foremost, I want to be free.

Given that, it was an easy choice for me. The FAANG company wanted a four-year, backloaded earn-out for the acquisition. That means you work there, and you get a little bit of the acquisition price every quarter over four years - with a smaller percentage at the beginning and a bigger chunk at the end (hence "backloaded").

On the other hand, the company that ultimately acquired Arcentry was happy to buy a license that gave them exclusive rights to the technology, along with a limited amount of integration support from my side and the option - but not the obligation - to provide more support in the future. All without any expectation of me working there.

They offered a little less than the FAANG, but for me, this was a much better deal.

Before bootstrapping your last business to an exit, you built another business with Venture Capital backers. How did the experiences differ?

It was like night and day. Not just because of the venture capital we've raised (1m from Blueyard Capital and another 750k in research grants and loans) - but also due to my lack of experience back then.

Deepstream started as a successful open source project, something I had some experience with from earlier things I've built, like golden-layout.com. And we were fairly focused on turning it into a modest but profitable company in the beginning.

But once the money came in, we started losing focus. Suddenly we had the time to polish it a bit more before going to market, to have a QA to test it, to have a PR agency to help with the launch. We've waited too long and invested too much into non-essential aspects.

In the end, we've spent eight months building out a realtime platform to compete with Google's Firebase, only to learn that we should have rather built an on-premise enterprise version. Turns out that the lean startup guy was right after all :-)

Now - this is by no means a problem with venture capital itself, just with me not being an experienced enough founder at the time.

With Arcentry I learned my lesson. Minimal costs - just me at home. Eight weeks from start to released (and charged for) MVP and iterating from there.

What is important to you in choosing the right investor to partner with?

Two things: Operational experience and a US network.

First: If you are a new founder, I feel it's crucial to choose an investor with operational experience. Someone, where at least one of the partners started and sold a successful company before. Unfortunately, in Berlin at least, this is a very hard thing to do - most VCs here have a purely financial background.

If you choose such a purely financial VC, make sure that you have some angels on board who have been founders before. Trust me; you need their expertise and network at least as much as the money itself.

Second: Choose a VC with ties to the US. Blueyard was great in this regard and opened the doors to Andreessen-Horowitz and similar firms in the states. What I learned in my (unsuccessful) quest of running up and down Sand Hill Road in Silicon Valley - as well as from talking to other founders - is that Venture Capital in the states is an entirely different ballgame - and the sooner you get into it, the better.

What is your plan for the future? Are you planning to launch another project?

I'm working on a new technology codenamed Hivekit - this time not as a solo venture but with a co-founder. It's still early days, but it will be a "spatial automation platform" - think of it as UI Path for the physical world.

You connect your vehicles, machines, people, and data sources and see them all live in a digital twin. From there, you can issue commands a la Command and Conquer, which the platform will translate into individual instructions.

What makes Hivekit really special, though, is that - rather than issuing commands yourself - you can automate them with a simple scripting language. You build up these atomic automation rules, and eventually, your factory, farm, scooter rental, taxi dispatch, logistics fleet, etc. will run as efficiently as a self orchestrating swarm.

‍