How to Install Pendo on a Single-Page App and Tag Your First Feature
Pendo combines in-app guides, product analytics and feedback for logged-in SaaS apps. Installing it takes an afternoon; making it useful takes a tagging plan. The free plan supports 500 monthly active users and paid contracts typically start near $7,000/year, so a half-configured install is an expensive mistake.
What does Pendo need before you start?
A user ID, an account ID, and a list of the three features you care most about. That is it — but the third item is the one teams skip, and it is the reason so many Pendo installs sit idle.
Pendo’s model is two-level: users belong to accounts. If your product is single-user with no team concept, map account ID to the user ID and accept that account-level analytics will be thin.
How do you install the Pendo agent?
- In Pendo, open Settings → Installation and copy your snippet (it contains your subscription key).
- Place it in the app shell, loaded on every authenticated route.
- On login, call
pendo.initialize()with the visitor and account objects. - On logout, call
pendo.clearSession()so the next user does not inherit the previous identity. - Confirm in the Pendo debugger that a visitor and account resolve.
Step 4 is skipped constantly. Shared devices and demo accounts end up merging two people’s sessions, and then the analytics team spends a week explaining an impossible retention curve.
pendo.initialize({
visitor: { id: "user_4821", email: "ada@example.com", role: "admin" },
account: { id: "acct_207", plan: "team", seats: 12 }
});
Why is tagging the part that matters?
Because Pendo’s analytics is built on features you define. Untagged, the tool knows pages and clicks but not meaning. Tagged, it can answer “which feature correlates with 90-day retention” — the question you presumably bought it for.
Tag by CSS selector inside the Pendo tagger, then name it in product language: billing.invoice.create, not .btn-primary-3. Names with dots or slashes group cleanly in reports; generic labels do not.
Tag three features in the first pass. Not thirty. Three features, one month of data, then decide whether to expand.
How do you avoid breaking tags on every release?
You will break some. Pendo tags depend on selectors, and any front-end refactor that changes class names can orphan a tag silently.
Three habits that limit the damage:
- Tag by stable attributes (
data-testid,id) rather than class names. - Re-check the tag count weekly; a feature with zero events after a release is a broken tag, not a dead feature.
- Keep a list of tag-to-selector mappings in your repo so the next engineer can see what depends on what.
This is the maintenance cost nobody prices in when comparing Pendo against lighter tools like Userpilot. Pendo is more capable and more work.
What should you do in the first week?
Use the free 500 MAU window to validate the data pipeline, not to evaluate features. Specifically: confirm events arrive within seconds, confirm user and account attributes populate, and confirm one guide can target a tagged feature.
If any of those fail, fix them before signing. A contract does not make the pipeline work.
Also run a cheap comparison: if you already pay for Amplitude or PostHog, check what Pendo tells you that they do not. Sometimes the honest answer is nothing, and the guide builder alone can be bought cheaper.
Where does Pendo get in the way?
It is not a weekend install. Expect two to three weeks of engineering for a full tagging pass, plus ongoing upkeep every time the UI changes. Teams without someone who owns product data will let it rot.
Guide publishing is genuinely good — no-code, well-targeted — but it is not a reason on its own to pay enterprise pricing. Buy Pendo for analytics, and treat the guides as the bonus.
Key Takeaways
- Pendo is free to 500 MAU; paid contracts commonly start near $7,000/year.
- Call
pendo.initialize()on login andpendo.clearSession()on logout, or user identities merge. - Tag features with stable selectors and product-language names; three features in month one is plenty.
- Weekly tag health checks catch silent breakage after front-end releases.
- If Amplitude or PostHog already covers your analytics, evaluate Pendo on guides alone.
