So long as you're making a loss, there's a lot of ways to "creatively" grow revenue. In the limit case (which would be fraud), you and I can just trade out expensive invoices, declaring each as a loss. This puts a lot of revenue on our books, with no value exchanged.

Mongo don't have to be committing fraud, of course. They could be doing any number of actual business activities that make legitimate trades --- but trades optimised at revenue growth, not profitability.

The major metric for companies used to be profit growth. When companies were optimising for that, it was smart to look at revenue growth as a leading indicator. But Goodhart's law ruins everything: now companies know to optimise for revenue growth, and so its value as a metric is much diminished.

Companies optimize for future cashflow, because that’s how value is created.

Mongo's current cashflow is negative, and the trend is increasingly negative.

I had a look at their prospectus, which describes the bulk of their revenue as subscriptions. Subscriptions sound like good unit economics. It's instructive to compare their pitch to investors to the pitch they make their customers: http://s3.amazonaws.com/info-mongodb-com/TCO_MongoDB_vs._Ora...

When Mongo talk to their customers, they describe the license cost as $0 --- they fold that into support. That sounds more like a service.

In other words: customer fires $100k of staff, pays Mongo $80k in "subscription", Mongo hires $120k of staff, which they tally up as "customer success". There's no expense category for support in their prospectus, so clearly the support personnel are filed under "sales and marketing".

It's the same old story. They're just selling $1 bills for $0.80 a piece.