Between 2013 and 2023, the operating stack of the median mid-market company roughly tripled in cardinality. The number of SaaS vendors a fifty-person company depended on went from somewhere around seven to somewhere around twenty-three. Everybody knows this. Everybody has the Okta dashboard to prove it.
What caused it
Three trends compounded. Cloud infrastructure made it trivial for a specialist vendor to ship a credible product. Credit-card procurement removed the central buyer. And the playbook of "land one team, expand across the org" meant every team got to pick their own vendor. The result was predictable: vendor sprawl at the organizational level, tab sprawl at the operator level.
The decade of unbundling ends when the cost of remembering exceeds the cost of consolidating.
What makes the stack close
We think the consolidation is driven by three forces working in concert, not any one of them on its own.
- AI amplifies context, which means fragmented data becomes dramatically more expensive per decision.
- Finance teams woke up. The "we are spending how much on SaaS?" conversation is now quarterly, not annual.
- The operator has less time than ever. A tool that demands a new mental model is a tool that loses.
The response in the market has been "superapps", "platforms", and "control planes" — three different words for the same instinct. We think the correct word is "memory". The winning architecture is the one where the company’s memory is readable by every tool, not the one where every tool has its own copy.
What that looks like
A single event log, a shared object graph, a single identity layer, and a canvas that lets an operator move between records without moving between apps. Everything else — AI agents, reports, dashboards — is a view on that memory.
This essay will not convince you on its own. The demo will. Until then, the argument is short: the stack got as fragmented as it is going to get. The next ten years close it.