Is PagerDuty worth it for a 100–1,000 person engineering team?

The money, named

What you pay. PagerDuty publishes a free plan for up to 5 users, Professional at $25 per user per month billed monthly and $21 billed annually (a published saving of $48 per user per year), and Business at $49 monthly and $41 annually (a published saving of $96 per user per year). Enterprise is custom-priced and therefore absent from every figure here. Multiply the annual rates by your responder count and twelve months:

Three things fall straight out of that table. The Business step at 500 responders is $120,000 a year — more than a senior SRE, for a feature set you should be able to name in one sentence. Paying monthly rather than annually costs $48 per user per year on Professional ($24,000 at 500 responders) and $96 on Business ($48,000 at 500), which is the most expensive way there is to avoid a commitment you are going to make anyway. And every licence you hold for someone no longer on a rota is $252 a year on Professional, $492 on Business.

What it has to beat. PagerDuty's own on-page estimator assumes a $100 per hour engineering rate and a 59% reduction in downtime, citing a Forrester Total Economic Impact study it commissioned. We will use our own, more conservative rate rather than the vendor's, and let you see both:

Professional, 500 responders      500 x $21 x 12   =  $126,000 / year
$126,000 / $90 per hour ( = $180,000 / 2,000)      =    1,400 engineer-hours / year
1,400 hours / 500 responders                      =  2.8 hours per responder per year
(at PagerDuty's own published $100/hr assumption  =    1,260 hours / year)

Two-point-eight hours per responder per year is a low bar — one incident where the right person is paged in ninety seconds instead of found in forty minutes covers a large chunk of it. For an org with a real production estate this is not the line to agonise over. The line to agonise over is the 500 in the left column, and whether it should be 300.

What it can cost you. Two things, both structural. The first is licence drift: rota membership is fluid and licences are not, so the count only ever goes up unless someone audits it. The second is add-ons. AIOps, Stakeholder licences, Status Pages, Live Call Routing and additional PagerDuty Advance AI Actions are all sold on top of the plans, and PagerDuty publishes no per-unit price for them on the pricing page. They are real budget lines — we simply refuse to guess at them, and you should insist on written per-unit pricing before any of them lands in a renewal quote.

The reason, at this size

1. The meter is the rota, and the rota is smaller than the org.

This is the single most valuable sentence on the page. A 500-engineer company does not have 500 people who need to be paged. It has some number of on-call rotations — perhaps 30 teams of 6 — plus a handful of incident commanders and platform staff. Everyone else needs to see incidents, not be woken by them.

At $252 per licensed user per year on Professional, the difference between licensing 500 people and licensing the 300 who are actually on a rota is $50,400 a year, and it does not remove a single alert from a single phone. PagerDuty sells a Stakeholder licence type for people who need visibility rather than paging — it publishes no price for it, so ask, but ask specifically, because "how many of our licences are people who have never been paged?" is the highest-yield question in this renewal.

2. Business buys process maturity, and process maturity is a decision, not a purchase.

What the published feature split adds on Business is largely about running incidents at scale: custom fields, up to three custom incident types, an expanded set of triggers and actions on incident workflows, advanced admin features, advanced ITSM integrations, internal status pages, multi-year historical data access, and a larger external status-page subscriber allowance.

Those are all real, and an org running a mature major-incident process will use them. But at $120,000 a year for 500 responders, the honest test is whether you have somebody who owns the incident process — writes the incident types, maintains the workflows, runs the reviews. If incident management at your company is currently "whoever picks it up starts a Slack huddle", buying Business does not create the process; it charges you for the absence of one. Buy the process first, on Professional, then upgrade when the tooling is the constraint.

3. SSO on the entry plan removes the usual security-tax trap.

It is worth calling out where a vendor does the decent thing, because it changes the shape of the decision. PagerDuty's published feature list puts SSO on Professional. Several tools on your renewals list gate SAML behind their most expensive tier and turn a security review into a forced upgrade — Sentry's Business plan, for one, re-prices every event as well as adding SAML.

Here, a security requirement does not push you to Business. Which means if you find yourself on Business, you should be able to name the operational feature that put you there. If you cannot, that is $120,000 a year at 500 responders with no stated reason.

4. The annual-versus-monthly gap is pure margin, and it is published.

PagerDuty publishes the saving itself: $48 per user per year on Professional and $96 on Business for committing annually. At 500 responders that is $24,000 and $48,000 respectively — for a tool you are certainly not switching off mid-year. Orgs end up on monthly billing by accident (a self-serve trial that converted, a team that expensed it before procurement existed), not by decision, and then never revisit it.

Check which one you are on before you do anything else on this page. It is the fastest five-figure saving here and it changes nothing about how the product works.

5. Alert quality decides whether any of this was worth buying.

The failure mode at this size is not the invoice — it is 200 alerts a week that nobody can act on, an on-call rota that treats the phone as background noise, and a real incident lost in the stream. No plan tier fixes that. PagerDuty will happily deliver 200 useless pages a week on Professional or on Business.

So the work that makes the $126,000 worth it is upstream and free: every alert has an owner and a runbook, anything that cannot be acted on at 3am is a ticket rather than a page, and a monthly review kills the top five noisiest alerts. Do that before you consider AIOps to suppress the noise you could have deleted — buying a noise-reduction add-on for alerts nobody wanted is paying twice for the same mistake.

What to do instead of just renewing it

Four controls. The first two are an afternoon and are worth five figures a year at this size.

1. Audit licences against actual rota membership, quarterly.

Export the user list and the escalation policies, and flag every licensed user who has not been on a schedule or been paged in ninety days. At $252 per user per year on Professional, 100 stale licences is $25,200 a year. Ask PagerDuty for written pricing on Stakeholder licences for the people who only need visibility — there is no published rate, so it has to come from the quote.

2. Check whether you are billed monthly, and stop.

PagerDuty publishes the difference: $48 per user per year on Professional, $96 on Business. At 500 responders that is $24,000 or $48,000 a year for a commitment you have already effectively made.

3. Name the Business feature, or move back down.

Custom incident types, advanced ITSM integration, internal status pages, multi-year data. If someone owns them and uses them, fine — $120,000 at 500 responders buys a lot of incident maturity. If nobody can name one in a sentence, that is the largest single saving on this page.

4. Kill the five noisiest alerts every month before buying anything to suppress them.

Every alert gets an owner and a runbook; anything not actionable at 3am becomes a ticket. Do this for two quarters before you price AIOps or any noise-reduction add-on — PagerDuty publishes no price for those, and you may find you no longer need the conversation.