On Building & Growing a "Google for Audio and Video" SaaS

This week I'm interviewing Marin, the founder of Omnisearch, a "Google for Audio and Video" SaaS. They're currently at $1k/MRR and growing. Marin has grown Omnisearch using app stores, cold prospecting and web conferences (more details on that below).

Hello! What's your background, and what are you working on?

My name is Marin, and I’m a developer originally from Croatia. After graduating from college, I first worked at a Y Combinator alum called SingleStore. Afterwards got a job at Amazon, working on AWS where I helped scale S3, and Alexa, shipping the email integration feature. After that I decided it was time for a change, so I went into the startup world, first with a machine learning-powered news app (now defunct) and then with our current product, Omnisearch.

Omnisearch was founded in late 2020 by myself and my co-founder Matej. Matej was one of the earliest engineers at fast-growing Techstars alum company called Memgraph. He also built the evaluation system used in all programming contests in Croatia.

Omnisearch is a B2B search product for everything, by which I mean every file type - from audio, video, PDFs, Word docs, plain text, images, presentations and more. Our main focus is site search, meaning that customers integrate Omnisearch into their own sites to supercharge the search functionality they offer their end users.

What motivated you to get started with Omnisearch?

The idea initially came from something that had been nagging me at Amazon: searching video materials. You can imagine that given the complexity of the products I worked on, a lot of relevant information was located in training videos, and finding what I was looking for in those hours of materials was really tedious and unproductive. So it seemed like a good idea to solve the problem in a general way and make it available to companies. Matej and I really hashed out the idea in late 2020 and decided to go all-in.

What went into building the initial product?

The initial version of Omnisearch was called Caption and it was a simple audio/video search API which we built in a couple of weeks. It was geared towards developers who could use it to build rich audio/video search experiences into their site. We even built a cool demo project, a podcast search engine, which was featured on Product Hunt.

The product really evolved from there, mainly in response to customer needs. This is when we rebranded intoto Omnisearch and truly became “search for everything” - we added support for searching PDF and Word docs, plain text, images, and more. And most importantly, we moved away from offering just the API towards a more end-to-end solution.

How have you attracted users and grown Omnisearch?

A super important step in getting initial traction was joining the Thinkific Partnership Program. Thinkific is a public company that’s basically Shopify for online courses. We were among the earliest adopters of their app store and developed a full-fledged Thinkific integration. We were able to get our first customers through this channel and grow our MRR significantly.

I’d say that partnerships and bizdev deals of this sort are a very underrated way of getting initial traction, and I’d suggest anyone starting out in B2B to seriously consider it. In addition, we’re making substantial investments into SEO and are also experimenting with a “powered by Omnisearch” link.

At the end of the day, though, nothing can replace direct reach-outs to prospective customers, either via LinkedIn or email. We currently use HubSpot to stay on top of all the prospects we’re talking to. And for prospecting we use LinkedIn Sales Navigator, as well as AlternativeTo and plain Google search for companies that might find Omnisearch useful. We haven’t used any cold outreach automation tools. We simply focus on really high-quality, personalized messages to prospects based on research of their use cases. As they say, “Do things that don’t scale”, at least in the early stages.

Apart from that, we derived a lot of value from conferences such as the recent Web Summit in Lisbon. We found that talking to prospects live is super efficient and great for building trust. We took advantage of Web Summit’s awesome mobile app that helps you connect with other attendees even before the conference itself. Overall, we had over twenty targeted meetings with companies in our first two verticals, education and media.

What's your business model, and how have you grown your revenue?

We have a pretty straightforward SaaS model with two components: a monthly subscription fee, and an initial processing fee for customers with large libraries of materials. The revenue has grown organically since launching, with basically all of our new signups being inbound.

What are your goals for the future?

We currently have two features in the pipeline that will help our customers refine their search results, so we’re laser focused on shipping those. After that we’ll basically focus on converting a lot more customers in the edtech space and growing our MRR. Slightly longer-term, our goal is to start expanding into other verticals. In addition to this, we’re investing heavily into our proprietary search algorithms.

If you had to start over, what would you do differently?

It would have been great if we’d switched to EdTech as our initial vertical a couple of months sooner, as it would have made it a lot easier to get some early traction. We were also a bit too eager to take on customization work for prospects that ended up not converting. But honestly, it’s been a pretty awesome fourteen months so I wouldn’t change much.

Have you found anything particularly helpful or advantageous?

Talking to customers and prospects! While there’s recently been some talk in high places against that, I still find it great advice, especially for B2B companies. Even getting a “no” is informative since it forces you to reevaluate and tighten your value proposition. The caveat is to always look for common threads and not over-optimize for a single customer.

The other piece of advice is to launch and iterate fast. I’m definitely a big proponent of the Lean Startup methodology, with its focus on shipping MVPs and getting feedback early. The thing to note here is that 2021 is not 2013, and that customers expect a bit more polished SaaS products. But the main principles remain, and it’s a far more common mistake to launch too late than too early.

Where can we go to learn more?

Learn more by checking out our site. You can follow us on Twitter and LinkedIn. And we also have a YouTube channel where you can check out some awesome demos!

If you have any questions, feel free to AMA in the comment section below.