PostHog vs Segment: Product Analytics Choice in 2026

Key Takeaways

• PostHog is a product analytics suite with replay, feature flags, and experiments, priced by usage, and the first 1,000,000 analytics events each month are free.

• Twilio Segment is sold today as a customer data pipeline and, on a separate quote, as a customer data platform. Connections pricing is public. The full platform price is Quote-based.

• The Team plan on Twilio Segment starts at $120 per month for 10,000 monthly visitors. Extra visitors in the next band cost $12 per 1,000.

• A free analytics allowance is common in this category, so a $0 starting line is not, by itself, a reason to pick either vendor.

• Anonymous visitors count toward Twilio Segment's monthly tracked users. On PostHog, person profiles and account-level group analytics are separate add-ons with their own rates.

• Neither purchase writes the onboarding handoff. Someone still has to confirm the account is real before a success team treats it as approved.

A product analytics platform measures behavior inside the product. A customer data platform collects events once, resolves identity, and routes records to other tools. TL;DR: choose PostHog when the open question is what users do after signup, and choose Twilio Segment when the open question is how one event reaches the rest of the stack.

Start with the job, not the logo

The search "posthog vs segment" looks like a tie between two logos. It is a category choice. PostHog's published product list is product analytics, session replay, feature flags, experiments, surveys, error tracking, logs, a data warehouse, and related usage products. Twilio's customer-data pages describe Connections as the pipeline that collects and sends data, and the Customer Data Platform as Connections plus Unify plus Engage. Those are different jobs. A team that needs funnels and replay will not get them from a routing contract. A team that needs one event fan-out to a large destination list will not get that catalog from an analytics bill.

Nothing in the current price pages says either product was discontinued or folded into the other. Twilio sells the pipeline and the platform under the Twilio Segment name. PostHog still sells PostHog. Treat older blog posts that talk about "Segment" as the standalone brand as the same pipeline, then read the live Twilio price pages before you compare a quote.

Free analytics events: 1,000,000/month at $0.00005 per event for the next million, according to PostHog (2026). That allowance is large enough that a small B2B product can stay on the free band for analytics events alone. It does not pay for a customer data platform, and it does not include every add-on. The wider market also gives many tools a free path: Tools with a free tier: 58% of the 84 product analytics tools in one directory, 49 of 84, according to Toolradar (2026). Use that as context, not as a quality score. A free tier means you can start. It does not mean the paid shape will match your billing model.

If the shortlist is still "which analytics product," read a wider product analytics tools for SaaS pass after this page. This comparison stays on the two names in the search.

Who this is for

This page is for a product lead or a data lead at a B2B SaaS company who is close to a decision. You already know you need cleaner product behavior, or a single pipe into the tools that already run lifecycle and revenue, or both. You can name the event that matters, such as trial started or workspace created, and you can roughly say whether you are counting events or people.

The buying committee usually splits. Product and engineering want funnels, retention, replay, and a flag to ship a fix. Marketing ops and lifecycle want the same event in email, ads, and the warehouse without a second instrumentation project. Finance wants a unit it can forecast. If those three asks are still one sentence, split them before you sign. PostHog answers the product ask on a usage meter. Twilio Segment Connections answers the routing ask on a visitor meter. The Customer Data Platform answers identity, audiences, and journeys on a custom quote.

Red flags: you cannot say whether success is a funnel inside the product or a record in another tool; you want audiences and journeys but you are only comparing analytics screenshots; you have no owner for the tracking plan, so either contract will bill you for events nobody trusts.

A workable checklist before the meeting:

• Write the job in one line: measure behavior, route events, or resolve identity and launch journeys.

• Count the unit the vendor bills. Events for PostHog. Monthly tracked users for Twilio Segment Connections.

• List the destinations that must receive the event in the first quarter, not the destinations you might want later.

• Decide whether account-level reporting (the company, not only the user) is in scope.

• Name the human who may reject a trial account before it becomes an onboarding task.

How we evaluated these two products

We did not run a lab test, and this is not a ranking from a private scorecard. The rubric below is a buying weight for a B2B SaaS lead whose search is this pair. Weights sum to 100 percent. They favor a wrong-category purchase over small feature gaps, because picking a pipeline when you needed analytics wastes the implementation either way.

Job match asks which sentence you will say in the first review. "We can see the funnel and the session" is a PostHog sentence. "The event left once and arrived in the warehouse and the lifecycle tool" is a Connections sentence. "We have a resolved profile and a journey" is a Customer Data Platform sentence, and that price is not on the public card.

Billing predictability asks whether finance can replay the invoice from a number the product team already watches. PostHog's meter is events, recordings, flag requests, and add-ons, with a $0 base on pay-as-you-go and lower rates at higher volume. Twilio Segment's public meter is monthly tracked users, including anonymous visitors. A viral docs site can raise the Segment bill while product usage of the app stays flat. Autocapture can raise the PostHog bill while logged-in accounts stay flat. Those are different surprises.

Identity model asks whether the company is the unit of analysis. PostHog can attach events to people and, with the group analytics add-on, to a company or workspace. Twilio Segment keeps identity resolution off the Free and Team connection plans and places it on CDP plans. If the board metric is "activated workspaces," say so before you pick a user-level pipe.

Warehouse path asks where the durable copy lives. PostHog sells its own warehouse product and priced exports. Twilio Segment includes a warehouse destination, with sync frequency and warehouse count tied to the plan. Implementation ownership asks who names events, who deletes test traffic, and who gets paged when a destination fails. Access and governance asks about seat caps, SSO, and whether schema control is included or sold as Protocols.

Our read of the weights: do not average the two products into one score. A 25 percent job-match miss is a disqualifier, not a tie-break.

What the public feature lists actually cover

The matrix uses the vendor pages opened for this guide. "Included" means the current price or plan page lists the capability on that plan. "Add-on" means the page lists it as optional or priced apart from the base unit. "Not this product" means the plan comparison does not sell that job. It is not a claim that a clever team could never move data there.

PostHog's price list describes autocapture of clicks and pageviews, web analytics included with product analytics, and no per-seat charge, so the whole team can use the project. The capture reference also tells implementers to send events to US cloud, EU cloud, or a self-hosted domain. That is a real hosting fork. Self-hosting still means your team runs the instance. Twilio Segment's published plans are cloud subscriptions with a free Connections tier, a Team tier, a custom Business tier, and a custom platform.

Two limits matter more than a logo row. First, identity resolution on the Connections comparison is marked available with CDP plans, not with Free or Team. Second, replay, flags, and product analytics are PostHog products. They do not appear as Connections features. A team that wants both jobs should budget both, or drop one job. Stacking them is coherent. Pretending one invoice covers both is not.

Our read: the disqualifier for PostHog is a requirement for governed fan-out into a large marketing and messaging catalog, plus journeys on resolved profiles. The disqualifier for Twilio Segment Connections is a requirement that the same login produce funnels, replay, and experiments without another analytics vendor. The disqualifier for the CDP quote is a team that only needed the pipe and does not have a journey owner.

Price list after the October check

Pricing checked October 10, 2026. The figures in this section are the vendor rates we read, not an estimate. Where Twilio does not print a platform price, the cell is Quote-based.

PostHog's Free plan is $0 and includes 1 project, 1 year of data retention, and community support. The pay-as-you-go plan has a $0 monthly base, 6 projects, 7-year retention, and email support. The first 1,000,000 product-analytics events each month are free, events 1,000,001 through 2,000,000 are $0.00005 each, and events 2,000,001 through 15,000,000 are $0.0000343 each, with the rate falling to $0.000009 above 250,000,000 events. The first 5,000 session recordings are free, then recordings through 15,000 are $0.005 each. The first 1,000,000 feature-flag requests are free, then requests through 2,000,000 are $0.0001 each, and experiments bill with flags. The first 1,000,000 warehouse rows are free, then rows through 10,000,000 are $0.000015 each. Group analytics starts at $0.000071 per event after 1,000,000 free events. The identified-events add-on starts at $0.000198 per event after 1,000,000 free events. The data-pipelines add-on starts at $0.000062 per event after 1,000,000 free events. The Boost package is $250 per month, the Scale package is $750 per month, and the Enterprise package is $2,000 per month, on top of usage. Eligible startups can apply for $50,000 in credits for 12 months if the company is under 2 years old and has raised under $5,000,000. Annual credit purchases from $25,000 to $59,999 are listed at a 20 percent discount. All of those rates are according to PostHog (2026).

Team starting price: $120/month for 10,000 visitors on Twilio Segment. The Free plan is $0 per month for 1,000 visitors, 2 sources, 500,000 reverse ETL records, 1 warehouse destination, and 700+ integrations, and the page intro also says 550+ pre-built destinations. The Team plan adds unlimited sources, 1,000,000 reverse ETL records, public API access, and a 14-day trial, with up to 20 percent off for annual billing. Extra monthly tracked users cost $12 per 1,000 in the 10,000 to 25,000 band, $11 per 1,000 in the 25,000 to 100,000 band, and $10 per 1,000 above 100,000. Free and Team include 10 seats and 50 hours of functions. Warehouse sync is once a day on Free and twice a day on Team. Business is custom pricing, with unlimited seats, unlimited warehouses, and a custom sync schedule. Those Connections figures are according to Twilio (2026).

The Business plan on Twilio Segment is Quote-based. The Customer Data Platform is Quote-based. Twilio's platform page says pricing is arranged with sales, that the bundle is Connections plus Unify plus Engage, and that Protocols stays an add-on. Packages on PostHog do not replace the usage meter. A $250 Boost fee still sits on top of events past the free band. The $0 in the Free row for Twilio Segment is a cap at 1,000 visitors, not an overage rate. Past that cap the published step is the Team plan, not a penny-per-visitor slide on the free tier.

Our read of total cost: forecast PostHog with events, and add group analytics and identified events if you will analyze known users and accounts. Forecast Twilio Segment with visitors, and remember anonymous traffic counts. Do not convert an event total into visitors with a guessed multiplier. If you only have one of those counts, the missing one stays a question for the vendor, not a made-up ratio.

A month of math on published rates

Take an illustrative B2B trial product that records 3,000,000 product-analytics events in a month and 12,000 monthly visitors in a pipe, using only the published bands and not a real customer's invoice. The first 1,000,000 PostHog events are $0, the next 1,000,000 at $0.00005 cost $50, and the third 1,000,000 at $0.0000343 cost $34.30, so base analytics is $84.30. The Team plan's $120 covers 10,000 visitors, and the extra 2,000 visitors at $12 per 1,000 cost $24, so the pipe is $144. The two lines add to $228.30. Every event in that sketch still needs a distinct_id, the capture field PostHog requires on the event, and the $84.30 uses only the base event rate, not the identified-events add-on and not replay. This is arithmetic on a fictional volume so a buyer can see the units. It is not a deployment, not a test, and not a promise that your month will land on these counts.

Read the combined line only if you truly need both jobs. If the funnel is the decision, the $84.30 analytics line is the relevant public math and the pipe is optional. If the decision is fan-out to tools you already run, the $144 pipe is the relevant public math and PostHog is optional. Replay, flags, group analytics, identified events, a Boost or Scale package, a second warehouse sync, or a CDP quote sit outside this sketch. Add them only after you know you will use them.

PostHog in practice

PostHog fits a product and engineering pair that will live in funnels, retention, paths, and replay, and that wants flags and experiments in the same project. A B2B team that thinks in workspaces should plan on the group analytics add-on instead of assuming the base event price covers account-level charts. The same team should decide early which events are allowed to create person profiles. The identified-events rate is a different meter from the base event rate, and turning it on for every autocaptured click will not look like the $84.30 sketch.

Marketplace score: 4.5/5 from 1,044 ratings, of which 69 percent are five stars, according to AWS Marketplace (2026). That page labels the ratings as external reviews from G2. It is a satisfaction signal, not proof that PostHog will match a CDP shortlist. Reviews on that page describe a real learning curve around event names, properties, and which report to open first. Treat that as an implementation warning. The product will record a lot, quickly, including through autocapture. A B2B team still has to decide which events define activation.

Limitations are specific. PostHog is not the Connections catalog, and the pipelines add-on is a usage meter rather than a substitute for a CDP quote. Journeys, generative audiences, and Twilio's profile API are not on the PostHog price list. Self-hosting is documented, and it moves ops work onto you. One project and one year of retention on the free plan will pinch a team that needs a separate production and staging project or a longer history. Pay-as-you-go raises those limits and still bills usage past the free bands. Spike traffic is your bill unless you set a billing limit in the product.

Implementation, from the public capture reference and the price list, starts with a project token and events that include an event name and a distinct ID. Autocapture can collect pageviews and clicks before the tracking plan is finished, which is useful for a first look and risky as a permanent design. A practical sequence is to define the few events that match the board metric, decide which of them may identify a person, add group keys if the account is the unit, and only then turn on replay for the flows that are hard to see in a funnel. Warehouse sync and batch export are separate usage lines. Budget a person to review event names weekly for the first month. Nobody on the price list does that naming for you.

Best fit is a SaaS product team that will make the next release decision from product evidence and can tolerate a usage invoice. Poor fit is a lifecycle team whose success criterion is "the event arrived in eight downstream tools and a journey started." That team should be on the Segment side of this page, or on both, with eyes open.

Twilio Segment in practice

Twilio Segment Connections fits a data or marketing-ops owner who must collect once and deliver to tools the company already pays for. The Free plan is a real $0 plan with hard limits: 1,000 visitors, 2 sources, and 1 warehouse. The Team plan is the first public paid step, at the $120 start above, and it is the plan that lists public API access. Business is the step for custom volume, governance, HIPAA eligibility with a BAA, and unlimited seats. The platform page is the step for identity-resolved profiles, audiences, and journeys.

Review score: 4.5/5 from 567 reviews, according to G2 (2026). Companies claimed: 25,000+ on the platform marketing line, and the Customer Data Platform price is customized rather than printed, according to Twilio (2026). The company count is Twilio's claim, not a census we repeated. The useful fact for a buyer is the split between a printed Connections price and a Quote-based platform. Identity resolution, the profile API, profile sync, traits, audiences, and journeys show up on the platform column. Protocols, the schema and quality control, is an add-on on both the Business pipeline and the platform comparison. AI predictions and recommendations are included in the platform description and sold as add-ons on the pipeline.

Limitations follow from that split. Connections will not become your funnel tool because you connected a source. Free and Team do not include identity resolution. Anonymous visitors count as monthly tracked users for the billing month, so a content site attached to the same workspace can move you up the $12, $11, and $10 bands without a matching rise in paid accounts. Ten seats on Free and Team will annoy a larger lifecycle squad. Syncing a warehouse once a day on Free, or twice on Team, is a poor match for a workflow that must react in minutes. Functions time is listed at 50 hours on Free and Team and custom on Business. Destination filters and user deletion exist as data controls on the comparison, but a full tracking-plan enforcement story is the Protocols add-on, not the free pipe.

Implementation is source install, event names, and a destination list you can defend. Start with the two sources that hold the trial event, one warehouse, and the one lifecycle tool that will actually act. Add destinations after a week of clean deliveries, not on day one. If the public API is required for a homegrown pull, you are already off the Free plan. If you need resolved profiles before a journey, stop the Connections-only evaluation and ask for the CDP quote. Compare that quote with the cost of keeping analytics in PostHog and routing only the approved events. Many teams do not need both on day one.

Best fit for Connections is a known visitor band and a destination list. Best fit for the platform is a team that will use audiences and journeys and can wait for a quote. Poor fit is a product manager who wanted replay of a broken signup and would still need a second tool after this contract. If the downstream tool is email and messaging, a separate look at Customer.io alternatives belongs next to the destination decision, not inside the CDP quote.

Owning the handoff in-house

The painful gap is usually not the chart. A trial event exists, and the success team still works from a spreadsheet. That gap is where teams start stitching Zapier, Make, or n8n, or a small internal service, between the analytics or pipeline tool and the system of record. Those tools can keep a run history, retry a failed step, branch on an error, and store an audit trail when someone configures them that way. The buyer owns the design. Observability, idempotency, escalation, access control, and maintenance do not appear because a connector exists. If two deliveries of the same account both create a task, that is an idempotency miss. If a failed write sits in a log nobody reads, that is an escalation miss. A homegrown worker can be the right answer when an engineer will own it. It is a poor answer when the design is a single unattended zap and the only alert is a full inbox.

A proposed configuration from US Tech Automations is aimed at that handoff, not at replacing PostHog or Twilio Segment. The trigger is a captured product event that already carries the distinct ID and a plan property the team chose to trust. The action pulls that event through an export or API the company already has credentials for, writes a queue row, and stops. A named reviewer confirms the account is a real trial and not an employee test. The output is an approved account handoff a success owner can work. Prerequisites are the export or API scope, a written rule for which event names qualify, and the human review point before any downstream record is treated as final. This is a configurable design. It is not a live customer, and it is not a measured result.

That US Tech Automations configuration can sit on the pipe side as well. The trigger is a track call that has already landed in the Twilio Segment warehouse sync, or a row available from the public API on a plan that includes it. The action compares the plan trait to the billing record and holds the row when the names disagree. The output is either an approved onboarding checklist a person releases, or a rejected row with a reason. Prerequisites are the sync or API access, a field map for the plan name, and a human review point before any message sends. The same design does not decide whether PostHog or Twilio Segment is the system of record. You pick that first. The handoff only moves an approved fact into the next team. For the surrounding onboarding steps, see how SaaS teams automate customer onboarding and what customer success software for SaaS companies is expected to hold after the handoff.

A proposed US Tech Automations design differs from a single unattended scenario in where the review and the retry policy are required up front. The connector tools can do retries and branches. The difference to configure is a named reviewer, a duplicate check on the account key, and a stop when the billing plan and the event plan do not match. If your engineer will build that in n8n and will watch it, you may not need another system. If the scenario would send a message before anyone looks, the miss is the review point, not the brand of the router.

When NOT to use US Tech Automations: skip it when PostHog or Twilio Segment already delivers the one event to the one system that acts, and a person already checks failures; skip it when a single maintained scenario in Zapier, Make, or n8n covers that path and someone owns retries; skip it when the team has not yet chosen the system of record, because an extra layer on an undecided pipe adds a second design before the first choice is real.

Mistakes that show up after the contract

The first mistake is buying the familiar name for the other team's job. Product picks PostHog, then marketing asks where the audiences are. Data picks Twilio Segment, then product asks where the replay is. Write the job line before the meeting so this mismatch is obvious while the quote is still a draft.

The second mistake is forecasting the wrong unit. A PostHog forecast that ignores autocapture will undercount events. A Twilio Segment forecast that counts only logged-in users will undercount monthly tracked users, because anonymous visitors count. A forecast that ignores the identified-events add-on will not match an invoice once person profiles are on. Use the bands on the price list. Do not borrow a multiplier from a blog.

The third mistake is turning on every destination or every recording in week one. Connections will happily fan out a messy event. Replay will happily store sessions you never watch. Start with the activation event, one warehouse, and one team that will act. Expand after names are stable.

The fourth mistake is skipping the review on internal traffic. Test accounts and employee clicks will create funnels and, on Twilio Segment, can count as visitors. Filter them on purpose. The fifth mistake is treating Protocols, group analytics, or a platform package as included because the base plan was affordable. Add-ons are how both vendors keep the starter price low. Read the add-on before you promise account-level reporting or a governed schema to the board.

Glossary

• Product analytics. Reports on what people do in the product, such as funnels, retention, and paths. This is PostHog's core job.

• Customer data platform. A system that collects customer events, resolves identity, and activates profiles in other tools. Twilio sells this as a Quote-based bundle.

• Connections. Twilio Segment's customer data pipeline. It collects and routes. It is not the analytics suite and not, by itself, the full platform.

• Monthly tracked user. Twilio's visitor unit for a billing month, including a unique anonymous visitor who never becomes a known user.

• Event. PostHog's analytics unit. A pageview, click, or custom action. Bands on the price list set the rate.

• Distinct ID. The identifier PostHog requires so events from the same actor can be tied together. Your team chooses the value and its quality.

• Identity resolution. Joining anonymous and known records into a profile. On these price pages, that job is on Twilio Segment CDP plans, not on Free or Team Connections.

• Reverse ETL. A sync from the warehouse back out to tools. Both vendors list warehouse-related syncs, with different limits and prices.

Questions buyers ask before signing

Is Twilio Segment a renamed or discontinued product?

No. The live price pages still sell Segment's pipeline and platform, under Twilio's name, and they do not describe a shutdown. Use those pages rather than an old brand screenshot when you compare a quote.

Is PostHog a customer data platform?

No. PostHog's price list is product analytics and the surrounding product tools, plus usage-priced pipes and exports. Unified profiles, audiences, and journeys, as Twilio describes them, are on the Quote-based platform, not on PostHog's event bands.

What does Twilio Segment charge for the first paid step?

The Team plan starts at $120 per month and includes 10,000 monthly visitors, with published per-1,000 rates after that. The Free plan is $0 and stops at 1,000 visitors. Business and the Customer Data Platform are Quote-based.

How far does PostHog's free analytics allowance go?

The first 1,000,000 product-analytics events each month are free, and the next band is $0.00005 per event. Replay, flags, identified events, and group analytics have their own allowances and rates, so a "free" project can still grow a bill once those are on.

Can one B2B team run both products?

Yes, when the jobs are actually different. PostHog can be the place you decide what the product did. Twilio Segment can be the place an approved event reaches the warehouse and the lifecycle tool. Pay both invoices only after you write down which system is allowed to define the account.

Do anonymous visitors change the Twilio Segment bill?

Yes. Twilio counts a monthly tracked user as any unique user it collects in the billing month, identified or not. A marketing site on the same workspace can move the invoice even when paid seats in the app do not.

When is a cheaper DIY handoff enough?

It is enough when one maintained flow already takes the approved event to the one system that acts, with retries, an error branch, and a person who looks at failures. It is not enough when failed rows have no owner, or when the flow messages customers before anyone confirms the account.

The decision after the shortlist

Choose PostHog when the next release depends on funnels, replay, flags, or experiments, and when you can forecast events, including the identified and group add-ons if accounts matter. Choose the Team or Business plan on Twilio Segment when the next release depends on one collection point and a destination list, and when you can forecast visitors, including anonymous ones. Ask for the Customer Data Platform quote only when identity resolution, audiences, and journeys are the actual scope. If the remaining work is the handoff from a trusted event to a person who owns onboarding, see how US Tech Automations configures this after the system of record is chosen, not before.

Re-read the unit on the invoice you are about to accept, then match it to the job line you wrote at the start. If they describe different problems, the lower starter price is the expensive choice.