Databases and MariaDB | Patrik Backman | Open Ocean | Neuvottelija 177

Summary

Patrik Backman of Open Ocean came on to talk about open source databases, how SaaS companies are valued, and early-stage investing. He was at MySQL from the early 2000s until 2008, and the best explanation in the episode is his analogy for what a database is at all: a bookshelf. Data is organised into books, the books go on the shelf, and the shelf gives the data back quickly.

MySQL. In the late 1990s Michael “Monty” Widenius and his co-founders put the database on the internet for free as open source, inspired by Linux, built to be as fast, reliable and efficient as the internet needed. The timing landed: all the early Amazons, Facebooks and YouTubes put their data into MySQL and pulled it back out to show anything on their pages. Backman joined a 50-person company and left a 600-person one with just under 100 million in revenue and an exit of roughly a billion.

Why MariaDB exists. MySQL was sold to Sun Microsystems in 2008; Sun did badly, and Oracle bought Sun — in Backman’s reading, precisely because MySQL was so popular. Customers, users and former employees doubted that a hard-edged commercial operator would keep an independent project free, and MariaDB was founded from the last stable open version. The argument for the fork is structural rather than ideological: Oracle has a commercial reason not to develop MySQL into a competitor for its own large systems, so it is constrained in where it can take the platform. Backman is candid that MySQL remains more popular; MariaDB is the alternative, around 20 million users and one of the ten largest databases in the world.

The SPAC. Miettinen raises an HS Vision analysis that had left the impression the listed company was close to worthless. Backman, no longer on the board and declining to speak for a listed company, explains the mechanism instead. The SPAC was priced a year before the listing, when markets were high. By the time the company reached the market, multiples had fallen, and the investors who had committed money redeemed it — entirely rationally. There was no dilution, because no shares were issued, but the funding did not arrive either. A valuation of about 700 million, a multiple of 14 on roughly 50 million of revenue, stood at 250 million and five times at the time of recording. Backman calls that a disappointment and far too low, arguing the price ignores the 20 million user base; Miettinen answers with scepticism about “eyeball valuations”.

What a buyer looks at. The ARR multiple, which compresses a complex valuation into one number and which Miettinen stresses is worth far more than ordinary revenue; Rule of 40, the sum of growth and profitability, below which a large part of the buyers walk away; Net Revenue Retention and churn, where five per cent churn is already very good; and gross margin, which small companies without a CFO routinely compute wrongly by leaving database costs out. Backman adds the investor’s view, which differs: a venture investment should return a hundredfold. His sharpest observation is that the first tenfold can be argued from the metrics, while the second — ten million to a hundred — depends on whether the team has the capacity and the ambition to hire people better than itself. That, he says, is the hardest judgement in venture capital.

The portfolio. Truecaller is the best of the cases because it is market-specific: it scaled in the Middle East and India because there the recipient pays for the call, so not knowing who is ringing costs money. The first version of Fonecta Caller was in fact licensed from Truecaller, which held every market except Finland. Supermetrics was profitable from day one and Open Ocean’s investment still sits untouched in its account; its origin was a Google representative promising a t-shirt to whoever could get Google Analytics data into Google Sheets, which Google itself could not then do. Nosto grew from five people and 100,000 in recurring revenue to 30 million. AppGyver was sold to SAP. MindsDB puts machine learning inside the database, and had just announced a Series A led by Benchmark.

ChatGPT in February 2023. Backman demonstrated it at the Christmas table, producing a thank-you speech in rhyming verse and extending it iteratively with the ham, three cats and the white wine. He also reports its failure: asked why Messi had never won a European championship, it explained that Argentina had not qualified for the final tournament. His advice to listeners is to found a software shop that codes with ChatGPT, delivers twice as fast and bids at three-quarters of the price.

Disclosure

The host states in the episode that he is himself a small investor in Leadoo, one of the portfolio companies discussed, and describes calling Leadoo’s largest customer as his own commercial due diligence before investing. He is also a partner at Translink Corporate Finance; Translink and Open Ocean share an address at Mikonkatu 1B in Helsinki, which the episode closes on.

Watch

The recording lives on the Neuvottelija channel: Tietokannat MariaDB | Patrik Backman | Open Ocean #neuvottelija 177. A Finnish edition of this episode is published at www.neuvottelija.fi.