PostHog vs Amplitude for B2B SaaS Leaders in 2026

Read the company before the user

A product or growth lead searching this comparison is close to a decision. The useful question is not which chart looks nicer. It is whether you can see behavior by company, predict the invoice, and still pass security review without guessing.

Product analytics for B2B SaaS is the record of what people do inside the product, read at the company level rather than only the user level. A user chart can say someone finished onboarding. A company view says whether that account did.

TL;DR: pick PostHog when you want a public rate card you can multiply, including a separate meter for company-level analysis. Pick Amplitude when a larger free event pool will carry the next stage and you will accept a custom quote for account reporting, single sign-on, and a higher replay cap.

Both products still sell under these names. Nothing on the pricing pages opened for this guide says either one was renamed, merged, or discontinued. This is the analytics decision, not a support-inbox decision. If the shortlist is still wider than these two, start with a wider look at product analytics tools for SaaS and then come back to the bill.

Free monthly replays in a comparison updated 29 September 2026 are 5,000 on PostHog and 10,000 on Amplitude, according to Growth Method (2026). Those two caps match the vendor pages cited below. Replay count is only one line on the invoice.

Key Takeaways

• Free analytics events: 1,000,000/month according to PostHog (2026). The same rate card sets the free plan at $0 with 1 project and 1 year of retention, sets pay-as-you-go at a $0 base with 6 projects and 7 years of retention and no per-seat fee, prices the next 1,000,000 analytics events at $0.00005 each and the next band at $0.0000343 each, prices group analytics at $0.000071 per event after 1,000,000 free events and at $0.00003 per event from 2,000,001 events through 15,000,000, prices an optional identified-events add-on at $0.000198 per event after 1,000,000 free events, includes 5,000 session recordings then $0.005 each, includes 1,000,000 feature-flag requests then $0.0001 each, includes 1,500 survey responses, lists the Boost package at $250 a month, the Scale package at $750 a month, and the Enterprise package at $2,000 a month, and offers eligible startups $50,000 in credits for 12 months when the company is under 2 years old and has raised under $5 million.

• Company-level reporting is not a free default on either product. PostHog sells it as a group add-on. Amplitude leaves group behavior analysis off the free plan.

• Amplitude publishes a free cap and a Plus plan that starts at $0. Growth and Enterprise are custom. You cannot finish a Plus overage in dollars from the public page alone.

• Review scores sit near each other. They are a satisfaction signal, not a feature verdict and not a price.

• The month that matters is the month company tracking is switched on, because that add-on can meter identified events that never received a company id.

• A chart does not assign the follow-up. If someone must review the account before a customer hears anything, that step sits outside both products.

Who this is for

This guide is for a product or growth lead at a B2B SaaS company who will live with the choice for more than a quarter. You have enough traffic to care about event volume, more than one person will read the data, and someone in finance or security will ask what happens at the next million events. You do not need a lab bake-off. You need the published limits and the places a quote replaces a rate.

Category average rating: 4.52/5 according to Toolradar (2026). The same page scores Amplitude at 4.50/5 from 3,059 reviews, in a set of 84 product-analytics tools of which 58% offer a free tier, with the 4.52 average taken across 61 rated tools. Use that only as context. A tenth of a point does not tell you whether company funnels are in the contract.

Red flags: events must stay inside a warehouse you already run and will not be copied into either product; you need SCIM, custom roles, and a named account manager before any paid usage; you need in-product guides and saved experiment results on a plan that does not go through a sales contract.

Our read of those flags is practical. PostHog's own product-analytics pricing page tells a buyer who wants data kept only in a warehouse such as Snowflake to pick another tool. Amplitude's plan grid puts experiment results, SSO, and the Accounts package on higher tiers. If one of the three flags is true, stop the trial and write the requirement down before anyone installs a snippet.

How we evaluated these tools

The weights below are buying criteria for this guide, not a score from a hands-on test. We did not run customer traffic through either product, and we did not invent a rank. A weight is how much of the decision that line should get for a B2B SaaS lead who is close to choosing. The numeric check is a published limit, a published price, or a count from the plan grid. Where a dollar rate is missing, the cell says so.

Company-level reporting carries the most weight because a B2B funnel that only counts users will misread accounts with many seats. A rate you can multiply comes next, because a custom renewal is a different risk from a usage bill you can cap. Replay, experiments, and admin controls split the rest. Seats are a small weight on purpose: both vendors already say the published plans do not charge per person.

Our read: if company reporting fails, do not buy the tool with the prettier retention chart. If the rate you can multiply fails, do not sign an annual commit from a slide that never states the next event. The tables that follow keep vendor facts in the cells and keep this weighting in the prose.

Published prices and the B2B meters

Pricing checked October 9, 2026.

Free events on Amplitude: 2,000,000/month according to Amplitude (2026). The same page lists the Free plan at $0 with no credit card, unlimited seats, unlimited feature flags, 10,000 monthly session replays, 1 year of data access, 1 month of replay retention, and 1 active feature experiment; lists the Plus plan as starting at $0 and scaling to 70,000,000 events, with 2 years of data access, 20 behavioral cohorts, 10,000 replays, and 1 active feature experiment; lists Growth at 20,000 replays and 2 active experiments and Enterprise at 50,000 replays and 2 active experiments, both with custom event pricing; marks group behavior analysis as not included on the Free plan; includes an EU data center option on every plan; and offers one free year of Growth to startups with under $10 million in funding and fewer than 20 employees.

Put the two free pools side by side before any demo. PostHog includes 1,000,000 analytics events. Amplitude includes 2,000,000. Amplitude also includes 10,000 replays against PostHog's 5,000. That gap favors Amplitude only while you stay inside the free plan and you do not need company-level reporting. The moment you turn on groups, PostHog adds a second meter. The moment you need Amplitude's Accounts package, the public page stops at a percentage of the platform plan and does not print the percentage.

The rate card lists the Enterprise package on PostHog at $2,000 a month. The packages page still sends that package through a conversation, so confirm the figure on the order form before you treat it as the invoice. Growth and Enterprise on Amplitude are Quote-based. The Plus plan on Amplitude is not Quote-based: it starts at $0, and the missing piece is the per-event dollar above the free events.

Our read of the grid: PostHog's usage lines stay calculable, and the admin packages stack on top as a flat monthly amount. Amplitude's free line is larger, and the B2B admin line is a quote. Neither pattern is a trick. They are different ways to buy the same job.

PostHog, priced in public

Best fit is an engineering-led B2B team that will calculate the bill, can live with one project on the free plan, and wants SQL against a warehouse that already has a published row price. The rate card calls the product open-source. Self-hosting is part of how reviewers describe it, which matters if security will not accept a vendor cloud yet. Startups under two years old that have raised under $5 million can ask about the $50,000 credit offer instead of guessing a discount.

Group types per project: 5 according to PostHog (2026). That page says group analytics is a paid add-on, billing starts when you enable it on the billing page, and the meter then includes all identified events in the project, not only events that already carry a company. Individual companies inside a type are not given a cap on that page. Anonymous events do not join groups, so a B2B funnel built only on anonymous traffic will look empty even after the add-on is on.

PostHog review score: 4.5/5 according to Toolradar (2026). The page aggregates 1,054 reviews and describes the product as open source and self-hostable. Read the cons on that page as setup effort and a bill that can grow, not as a broken free tier. The free tier is real. The growth in the bill is also real once identified events and groups are both on.

Limitations are specific. Group analytics is a second invoice line, not a checkbox inside the base event price. The optional identified-events add-on is a third line if you subscribe to it, starting at $0.000198 after 1,000,000 free events, and it is not included in the illustrative total later. Autocapture records clicks and pageviews without a tracking plan, which can spend the free million before a human event is defined. Boost at $250 a month is what you add for SSO enforcement and a HIPAA BAA. Scale at $750 a month is the SAML step. Those packages do not replace usage fees.

Implementation is a sequence, not a slogan. Create the project, send each event with a stable person id, and do not enable the group add-on until you have counted identified volume for a normal week. Set a billing limit on each product before a card is added. On the free plan, events past the allowance are dropped. On pay-as-you-go, the limit is the ceiling PostHog will charge. A person should look at that projected number before the limit is raised. Feature flags share the request meter with experiments, so a flag evaluated on every page view is not free just because the experiment count is small.

Amplitude, quoted past the free line

Best fit is a team that wants the larger free pool, a replay cap of 10,000 on day one, and guides inside the product, and that will move to a quoted plan when account reporting and SSO become requirements. The Free plan needs no card. The Plus plan needs a card even though it starts at $0. Growth and Enterprise need a sales conversation. If in-product guidance is the real reason you are here, read Pendo and Amplitude for product-led growth before you assume guides are a reason to skip the quote.

Free-plan block: after 3 overages according to Amplitude (2026). The same page says a Free plan that stays over the event limit can be deleted 6 months after the initial block, and the one-year startup offer includes 100,000,000 events or 200,000 MTUs, with a guardrail of 1,000 events per MTU on MTU-based Plus or Startups plans. An MTU is a person who triggers at least one event in the month. The guardrail does not apply to plans that bill on an event quota instead.

The plan grid marks group behavior analysis as not included on Free and available on Plus, Growth, and Enterprise. The Accounts package, which lists account-level analytics, group-level experiments, account bucketing, and B2B reporting, is shown for Growth and Enterprise and is priced as a percentage of the platform plan. That percentage is not printed. Do not treat the word Available on Plus as the same thing as the Accounts package.

Limitations follow from that split. You cannot compute the Plus overage from a public per-event rate. Experiment results are not on Free or Plus. Active experiments are 1 on Free and Plus and 2 on Growth and Enterprise unless you buy more. Replay retention on the published grid is 1 month unless an add-on extends it. Three bad months on the free plan do not merely warn you. They block charts, and a continued overrun can delete the account 6 months after that block. Startup eligibility is narrow: fewer than 20 employees and under $10 million in funding, for one year of Growth, not a permanent free enterprise tier.

Implementation means instrumenting on purpose. Autocapture is included on every plan, so a careless snippet can spend the 2,000,000 events. Decide the activation event for a company before you argue about charts. If you qualify for the startup year, ask what happens in month eleven, because the billing notes say the year is not extended and the fallback free plan returns to 2,000,000 events. Keep a person on the quota each month. A blocked workspace is a bad week to discover that finance never saw the cap.

An illustrative month at 4,000,000 events

Take an illustrative month, not a measured customer: the product sends 4,000,000 identified events, each with a distinct_id on the capture request, and group analytics is already enabled. Product analytics bills the first 1,000,000 at $0, the next 1,000,000 at $0.00005 for $50, and the remaining 2,000,000 at $0.0000343 for $68.60, so that line is $118.60. The group add-on meters the same identified events: 1,000,000 free, the next 1,000,000 at $0.000071 for $71, and the remaining 2,000,000 at $0.00003 for $60, so the add-on is $131. Together those two lines are $249.60. If SSO means the Boost package as well, add $250 and the illustrative total is $499.60. Recordings inside the 5,000 free recordings add $0. The optional identified-events add-on is not in this total. The same 4,000,000 events would clear Amplitude's free plan by 2,000,000 events. The Plus plan starts at $0 and can hold 70,000,000, but the public page does not print the per-event dollar, so that half of the comparison stays unfinished until a quote arrives.

Our read: $249.60 is the honest usage picture for that assumed month, and $499.60 is the picture if admin controls are required in the same month. It is not a promise about your traffic. Change the 4,000,000 and the bands change. Leaving group analytics off removes the $131. It also removes the company funnel you probably came here to buy.

What happens after the insight

The analytics choice still leaves a gap. Someone has to decide what the company should receive next, and that decision should not be an automatic email from a chart. A proposed workflow, not a live deployment, starts when a scheduled export shows a company crossed the activation event you defined. US Tech Automations can read that export, skip the company if a review item already exists, and open an internal item that names the account, the event count, and the suggested next step. Prerequisites are an export or API credential from the analytics product, a stable account id, and a place to store the review item. A person approves or rejects the item. Nothing customer-facing goes out before that approval.

A second proposed path watches the meter rather than the account. When captured events cross a threshold the team sets below the next paid band, US Tech Automations drafts an internal note with the count and the published limit, then waits. A person chooses to raise the billing limit, cut autocapture, or leave the cap alone. The workflow does not change the vendor plan and does not tell the customer their usage changed. The same human checkpoint belongs in any handoff into customer success software for B2B SaaS. If the number you are watching is a commercial meter rather than a product funnel, price that problem with usage metering software for B2B SaaS before you bury it inside an analytics invoice.

The reader's other path is to stitch the same handoff in Zapier, Make, or n8n, or to build it in-house. Those tools can keep a run history, retry a failed call, branch on errors, and store audit evidence when someone configures them to do so. The buyer then owns observability, idempotency, escalation, access control, and maintenance, including what happens when the analytics token expires. A proposed US Tech Automations design would configure the duplicate check, the approval stop, and the escalation path as named steps, with the analytics credential and the account system connected first, and with a person still required before any message leaves the building. That is a different operating shape, not a claim that a small automation cannot work.

Mistakes that make the bill a surprise

Comparing only the free event caps is the expensive mistake. Amplitude's 2,000,000 against PostHog's 1,000,000 looks like a clear win until group analytics bills the identified stream a second time. Price the month you will actually run, not the month before the add-on exists.

Turning the group add-on on during a launch week is the second mistake. Billing starts at enablement, and it counts identified events that have no company on them yet. Wait until the tracking plan marks which events are identified.

Reading Available on Plus as full B2B account reporting is the third. The Accounts package, with account-level reporting, is listed on Growth and Enterprise and priced as a percentage that is not public. Ask for that percentage in writing.

Letting autocapture run with no billing limit is the fourth. PostHog will drop events on the free plan or charge up to the limit you set. Amplitude can block the free workspace after the third overage. Either outcome is worse than a weekly look at the counter.

Treating 4.5 and 4.50 as a winner is the fifth. The review counts are large enough to take seriously and too close to end the purchase. Use them to check for setup pain, then go back to the rate card.

Buying an orchestration layer before the activation event exists is the sixth. A workflow cannot review an account the analytics product cannot name. Define the company id and the trigger event first.

Glossary

• Event: one recorded action, such as a page view, a click, or a named product action. Both vendors bill or cap around this unit.

• Identified event: an event tied to a person profile. On PostHog, anonymous events do not attach to groups, and the group add-on meters identified events once it is enabled.

• Group analytics: analysis by an entity such as a company. On PostHog it is a paid add-on with up to 5 group types in a project.

• MTU: a monthly tracked user on Amplitude, a person with at least one event in the month. MTU-based plans also carry a guardrail of 1,000 events per MTU.

• Session replay: a recording of a visit. The free pools are 5,000 recordings on PostHog and 10,000 on Amplitude's Free and Plus plans.

• Feature flag request: one evaluation of a flag. PostHog includes 1,000,000 a month and then charges. Amplitude lists flags as unlimited and meters experiments separately.

• Cohort: a saved set of users who share a behavior or trait. Amplitude's Plus plan includes 20 behavioral cohorts. A cohort is not the same thing as a company group.

• Billing limit: a ceiling you set so usage cannot charge past a number you chose. It is the practical control on a usage plan, and a person should own any change to it.

Questions to settle before you buy

Is PostHog cheaper than Amplitude for a B2B team?

PostHog is cheaper to forecast, and Amplitude is cheaper only while you stay inside 2,000,000 events and you do not need the Accounts package. At the illustrative 4,000,000 identified events with group analytics on, the PostHog usage lines add to $249.60 before any platform package. Amplitude's Free plan would already be 2,000,000 events over, and the public Plus page does not give the dollar rate for those events. Growth and Enterprise are Quote-based. If your real volume stays under 2,000,000 and company reporting can wait, Amplitude's free plan costs $0.

Does either free plan include company-level analytics?

No. Amplitude's Free plan marks group behavior analysis as not included, and PostHog's group analytics is a paid add-on even though the first 1,000,000 events on that meter are $0. You can explore user funnels on both free plans. You should not tell a board you have account-level activation until the add-on or the Accounts package is actually on the contract. Enabling the PostHog add-on also changes the bill for identified events that never got a company property.

What happens after three months over Amplitude's free cap?

Amplitude blocks the Free plan after the third month over the event limit, and it can delete the account 6 months after that block if you still have not upgraded. Charts stop being available while admin functions such as the user API can remain. That is a sharper cliff than a usage overage you can cap. If you are near 2,000,000 events, assign someone to read the counter weekly rather than discovering the block in a planning meeting.

Do platform packages replace usage fees on PostHog?

No. The Boost package is $250 a month, the Scale package is $750 a month, and the Enterprise package is listed at $2,000 a month, and analytics events are still metered. You buy the package for controls such as SSO, a HIPAA BAA, SAML, or RBAC. You still pay the event lines. Adding Boost to the illustrative usage total moves $249.60 to $499.60. Confirm the Enterprise list price on the order form, because the packages page also routes that tier through a person.

Which product should you choose for in-product guides?

Choose Amplitude if guides are a requirement, and expect the free allowance to be 1 guide or survey, not an unlimited onboarding program. PostHog includes surveys, with 1,500 responses free, and does not sell the same guide product on the pages opened for this guide. Experiment results on Amplitude arrive with Growth, not with the Free plan. If guides are the whole project, this analytics comparison is the wrong final filter.

When NOT to use US Tech Automations?

Skip the extra orchestration when the analytics product is only a chart nobody acts on, when one simple automation with a single owner already covers the handoff, or when a native cohort sync already writes the one list you need and a person checks it. Zapier, Make, or n8n are enough in those cases if you accept owning retries and access. A second system is wasted motion when the trigger is unclear or the volume is so low that a weekly review of the chart is the whole process. The proposed design earns its place only when the duplicate check, the approval stop, and the meter alert would otherwise be rebuilt and maintained by the same small team.

Choose the analytics product for the company view and the bill you can explain. Use a separate, review-gated workflow only after that choice is real. You can see how US Tech Automations configures this when the export, the account id, and the human approval step are the part you still have to design.