Skip to content
Notifications
Clear all

Unpopular opinion: Okta's documentation has gotten worse, not better.

2 Posts
2 Users
0 Reactions
3 Views
(@cost_optimizer_88)
Estimable Member
Joined: 3 months ago
Posts: 95
Topic starter   [#9014]

Let’s start with the premise that a product’s documentation is its silent support engineer. If that’s true, then Okta’s documentation has been on a quiet vacation for the past few years while the bill for its absence gets footed by every engineering team trying to integrate it. I’ve seen the common refrain that it’s “improving,” but when you actually track the time spent deciphering it versus building, the numbers tell a different story. This isn’t about nitpicking typos; it’s about the systemic degradation of a critical resource that directly impacts implementation cost and time.

Consider the evolution of a simple task: setting up a SAML app integration for a custom app. Three years ago, you might have had a linear guide. Today, you’re presented with:

* A labyrinth of “getting started” pages that all seem to circle back to each other.
* Code samples that are often deprecated or reference SDK versions that are no longer maintained, without clear warnings.
* Critical configuration nuances buried in community forum replies from 2021, which themselves are now outdated because the Admin UI has been redesigned twice since then.

The most egregious issue is the fragmentation. Information isn’t just in one place. You have the “official” docs, the “developer” docs, the “help” articles, and the API references, each with a different tone, detail level, and sometimes contradictory advice. It’s like a poorly managed multi-cloud strategy where you’re paying data egress fees just to find a complete answer.

Take their API documentation as a concrete example. You want to automate user provisioning. The API reference lists the parameters. What it frequently omits are the inter-dependencies between those parameters or the specific conditions under which certain fields are required. You end up in a trial-and-error loop, which from a cost perspective, is pure waste of developer hours.

```bash
# Example: Their API docs might show this for updating a user:
curl -X POST
"https://${yourOktaDomain}/api/v1/users/${userId}"
-H "Authorization: SSWS ${api_token}"
-H "Content-Type: application/json"
-d '{
"profile": {
"firstName": "New"
}
}'

# But silently, if you're using certain profile sourcing rules, this might be ignored or require additional, undocumented parameters in a different endpoint call. The failure mode is a 200 OK with no change, not a helpful error.
```

This isn’t an edge case. It’s a pattern. The documentation feels built for checkbox compliance (“we have docs”) rather than for enabling successful, efficient customer outcomes. For a product with Okta’s market position and price point, this is an indirect tax on its customers. Every hour a senior dev spends cross-referencing three articles and testing in a sandbox to understand a basic workflow is an hour not spent on core product work. In cloud terms, it’s a massive, unmonitored resource leak.

The counter-argument is that they’ve added more content, more guides, more videos. Volume ≠ quality. In fact, excessive, poorly organized information increases the cognitive load and search time. It’s the equivalent of over-provisioning instances without auto-scaling: you’re throwing more resources (pages) at the problem without making the system (knowledge retrieval) more efficient. The signal-to-noise ratio has plummeted.

I’d be less sardonic about it if this was a free tool. But when you’re allocating a significant portion of your SaaS budget to an identity platform, you’re paying for the whole ecosystem, including clear, reliable, and maintainable documentation. The current state suggests that investment isn’t being prioritized, and the cost is being externalized to every team trying to make it work.


pay for what you use, not what you reserve


   
Quote
(@jakes)
Estimable Member
Joined: 1 week ago
Posts: 74
 

Your point about tracking time spent vs building is the key. I've had teams waste days because of this. The real cost is in the telemetry gap: you can't measure how many hours are lost across all dev teams chasing bad docs, so the problem never gets priority. It's an invisible tax.

The fragmentation issue isn't just annoying, it breaks reproducible setups. If I can't give a new engineer a single doc path that works, my onboarding and deployment timelines are shot.

They keep adding features but the docs feel like an afterthought. A changelog tied to deprecation notices would be a start.


Show me the methodology.


   
ReplyQuote