GA4 vs Mixpanel vs Amplitude: Which Stack for a Lean Team

For a lean marketing team this is rarely a three-way choice. GA4 (Google Analytics 4) is free, probably already installed, and handles acquisition reporting — where traffic came from and what it cost. Mixpanel or Amplitude sits alongside it to answer what GA4 is bad at: what specific people do after they arrive, and whether they come back.

So the real decision is which of those two you add. And for most small teams it is settled by one thing: the meter. Mixpanel bills you on events. Amplitude bills you on monthly tracked users. The same website can sit comfortably inside one free plan and blow straight past the other, so the arithmetic below matters more than any feature comparison.

The short version

If you look like this	Add this	Why
High traffic, light engagement (content site, D2C store, lead-gen site)	Mixpanel	Lots of people firing few events each. The user-based meter prices you out first.
Low traffic, heavy engagement (SaaS app, dashboard, marketplace seller tool)	Amplitude	Few people firing hundreds of events each. The event meter burns out first.
Your product or data team already runs one of them	Whichever they run	A shared event taxonomy beats a marginally better tool in a second silo.
Hundreds of thousands of anonymous visitors, zero budget	GA4 plus BigQuery, or PostHog	Both commercial free tiers break at that volume unless you filter hard.
Your real gap is messaging, not analysis	CleverTap, MoEngage or WebEngage	Analytics plus push, email and WhatsApp in one place.

The free tiers, side by side

Two definitions first, because the whole decision rests on them. An event is one tracked action — a page view, an add-to-cart, a signup. An MTU (monthly tracked user) is one identified person seen at least once in a calendar month, no matter whether they fire three events or three thousand.

	GA4 (standard)	Mixpanel (free)	Amplitude (Starter, free)
Built for	Web and app acquisition, channels, campaigns	Product and funnel analytics	Product and behavioural analytics
Metered on	Nothing on the free tier	Events per month	Monthly tracked users
Published free allowance	No monthly event charge; soft collection limits instead	About 1 million events/month (cut from 20 million in 2024)	About 50,000 MTUs
Limits that bite in practice	500 distinct event names, 50 event-scoped custom dimensions, sampling in Explorations on big queries	Event ceiling; advanced modelling and some integrations gated to paid	MTU ceiling; limited seats and saved charts
Raw data export	BigQuery export, free up to roughly 1 million events/day	Generally paid tiers	Generally paid tiers
Data retention	User- and event-level data: 2 or 14 months	Per plan terms	Per plan terms
First paid step	GA4 360, commonly reported from around $50,000/year	From roughly $20-30/month, scaling with events	From roughly $49/month, scaling with MTUs

Treat every number in that table as a starting point with a shelf life. All three vendors have changed their free tiers more than once, and Mixpanel’s 20x cut in 2024 caught a lot of small teams mid-quarter. Before you commit, open Mixpanel’s pricing page, Amplitude’s pricing page and Google’s GA4 limits documentation and check the current figures yourself. Anyone quoting these from memory, including an article, is quoting a snapshot.

Web analytics vs product analytics: the actual difference

Web analytics is session-shaped. It answers: how many sessions, from which channel, landing on which page, at what cost per conversion. GA4 is the dominant free tool here and it wires into Google Ads, Search Console and BigQuery without you writing code.

Product analytics is person-shaped. It answers: of the people who signed up in March, how many were still active in June, and which first-week action predicted it. Mixpanel and Amplitude are built around that — cohorts, retention curves, and funnels you can re-slice after the fact without re-tagging anything.

The practical split: GA4 tells you which ad to switch off. Product analytics tells you which onboarding step to rebuild. A lean team almost always needs both answers, which is why “GA4 versus Mixpanel” is the wrong framing. It is GA4 plus one of them.

Do the volume maths before you pick

This takes ten minutes and saves a quarter of regret. You need two numbers, not one:

1. Monthly events = monthly users × sessions per user × tracked actions per session
2. Monthly tracked users = the count of distinct identified people you will send

Then compare each against the relevant free ceiling. Three worked examples, arithmetic shown.

Example 1 — a D2C store. 70,000 monthly visitors, 1.3 sessions each, 10 tracked actions per session (page view, product view, variant select, add to cart, three checkout steps, purchase, and a couple of scroll or search events).

• Events: 70,000 × 1.3 × 10 = 910,000 events/month — 91% of a 1 million allowance, with zero festive headroom.
• MTUs: 70,000 — already 40% over a 50,000 ceiling, so you are paying from month one.
• Now add a Diwali spike at 2.5x traffic: 2.28 million events. Both meters break, but the event-metered bill at 2.3 million events is a far smaller step than the jump from free to a 175,000-MTU plan.

Verdict: the event meter, and plan for the spike month in the budget rather than discovering it on the invoice.

Example 2 — a B2B SaaS product. 6,000 monthly active users, 45 events each.

• Events: 6,000 × 45 = 270,000 events/month — about a quarter of a 1 million allowance.
• MTUs: 6,000 — about 12% of a 50,000 ceiling.

Verdict: comfortably free on either meter. Pick on usability and on whichever one your product manager will actually open. But run the projection forward: at 3x growth in events-per-user — normal as a product adds features — 6,000 users firing 150 events each is 900,000 events, and the event meter is suddenly the tight one while MTUs are still nowhere near the cap.

Example 3 — a content and media site. 400,000 monthly readers, 1.2 sessions, 4 events per session.

• Events: 400,000 × 1.2 × 4 = 1.92 million events/month.
• MTUs: 400,000.

Verdict: both free tiers are gone, and the paid tiers at 400,000 MTUs are not lean-team money. The move here is not to pick a tool but to reduce what you send: keep GA4 with the BigQuery export for the anonymous audience, and send only identified users — newsletter subscribers, logged-in readers, purchasers — to the product analytics tool. A media site with 400,000 readers and 25,000 registered ones is a 25,000-MTU account, which is free.

Four levers that keep you inside a free tier

Before you accept the invoice, check whether you can shrink the input:

• Send identified users only. The single biggest lever on an MTU meter, as above. Anonymous traffic belongs in GA4.
• Stop sending decorative events. Scroll depth at 25/50/75/100% is four events per page view and almost never changes a decision. Dropping it can cut 30-40% of volume on a content-heavy site.
• Filter server-side. If you route tags through a server container you can drop bot traffic, internal IPs and junk events before they count against anything. Our guide to server-side tagging in GTM covers the setup without the jargon.
• Collapse near-duplicate events into properties. One “Product Viewed” event with a category property beats eight category-specific events, and it keeps you well away from GA4’s 500-event-name ceiling too.

Mixpanel vs Amplitude in practice

Feature lists have largely converged. Both do funnels, retention, cohorts, session replay as an add-on and warehouse connections. The differences that matter to a five-person team are narrower than the marketing pages suggest:

• Learning curve. Mixpanel is generally faster for a non-analyst marketer to self-serve on. Amplitude has more analytical depth and correspondingly more surface area to get lost in.
• Who else uses it. Amplitude is more often the incumbent in product and data teams. If your PM already has it, adding marketing events there beats standing up a parallel tool and a parallel truth.
• Seats on the free plan. Check the current seat count. A tool only two people can log into stops being the team’s source of truth inside a quarter, and “ask Priya to pull it” is not an analytics practice.
• The meter, again. It will decide your bill far more than any feature will.

If neither free tier fits and the budget is genuinely zero, look at PostHog, which publishes a monthly free event allowance and can be self-hosted — at the cost of engineering time a lean team usually does not have spare.

What it costs in rupees

Mixpanel and Amplitude bill in US dollars, so your rupee cost moves with the exchange rate and normally attracts 18% GST on top as an import of services. Check the live rate, but at roughly ₹88 to the dollar the arithmetic looks like this:

Plan	USD/month	Approx. INR/month	With 18% GST	Approx. INR/year
Mixpanel, entry paid	~$25	~₹2,200	~₹2,600	~₹31,000
Amplitude, entry paid	~$49	~₹4,300	~₹5,100	~₹61,000
GA4 360	~$4,200	~₹3.7 lakh	~₹4.4 lakh	~₹44 lakh

Two honest caveats. First, the entry paid tiers are floors, not what you will pay — both scale with volume, so model your own usage. Second, GA4 360 pricing is negotiated and not publicly listed; the figure above is the commonly reported starting commitment and should be treated as an order of magnitude, not a quote. For the wider picture of what a full stack costs an Indian team, see our Indian marketing stack cost reference.

The India angle

Indian teams have a third path worth pricing. CleverTap, MoEngage and WebEngage bundle behavioural analytics with campaign execution — push, WhatsApp, email, in-app. If your real gap is “we can see the drop-off but can’t message the people who dropped off,” one of those can replace both a product analytics tool and a separate messaging stack. They quote per monthly active user and almost always require a sales conversation, so budget materially more than $49 a month — you are buying messaging infrastructure, not just charts. Our comparison of the three goes through where each one fits. If you only need analysis, they are heavier than you need.

A second India-specific point that costs teams real money: a large share of Indian D2C traffic is app-and-web mixed, often on low-end Android devices on patchy networks. Test that your event collection survives a bad connection before you trust any funnel number. Dropped events look exactly like genuine drop-off in all three tools, and a 12% “checkout abandonment problem” that is actually a payload failure will send you rebuilding a checkout that works fine.

Third: with consent banners and cookie loss now normal, the identity layer matters more than the tool. If you are not yet collecting email or phone at a point where you can legitimately stitch sessions together, read our first-party data strategy guide before you spend a month instrumenting events you cannot attribute to anyone.

Failure modes lean teams actually hit

• Tracking everything on day one. Every decorative click event eats the free allowance and pushes you toward GA4’s 500-event-name ceiling. Start with 15-25 events that map to decisions you actually make.
• Naming events on the fly. “Add to Cart”, “add_to_cart” and “AddToCart” become three unrelated events and nobody notices for five months. Write a one-page tracking plan before anyone instruments anything.
• Leaving the GA4 BigQuery export off. GA4’s standard tier keeps user- and event-level data for 14 months at most. Enable the export now and you own your history; enable it in year two and year one is simply gone.
• Expecting GA4 to do real retention analysis. It will approximate it. It will not give you the cohort view a product analytics tool gives you, and forcing it burns days you do not have.
• Discovering the overage during a sale. Set a volume alert at 70% of your allowance. Festive-season traffic is exactly when you least want a sampling notice or a surprise invoice.
• Paying before a decision is waiting on the data. If nobody has asked a question the free tier cannot answer, the upgrade is procurement theatre.
• Reconciling to the last order. Blocked cookies and consent refusals mean all three tools undercount. Compare trends, and never try to tie analytics to your payment gateway order-for-order.

A 30-day rollout that works

1. Week 1: write the tracking plan — 15-25 events, one naming convention, properties listed, owner named.
2. Week 1: turn on the GA4 BigQuery export. It costs nothing at small volumes.
3. Week 2: do the volume maths from both meters, including a spike month, and pick the tool.
4. Week 2-3: instrument the plan, identified users first. Verify on a throttled mobile connection, not your office wifi.
5. Week 3: build exactly three things — one acquisition-to-signup funnel, one retention curve, one activation cohort. Nothing else.
6. Week 4: set the 70% volume alert, give every teammate a login, and put the three charts in a recurring meeting. A dashboard nobody opens is worse than no dashboard, because it looks like coverage.

What this means for you

• Keep GA4 regardless. It is free, it is the shared language with agencies and ad platforms, and there is no upside to removing it.
• Do the volume maths first. Users × sessions × events per session, and separately your identified user count. Those two numbers pick your tool faster than any feature matrix.
• High traffic, light engagement? The event meter is usually the cheaper free ride. Low traffic, heavy engagement? The user meter is. Under about 10,000 engaged users, either works — pick on usability.
• Shrink the input before you accept the bill. Identified users only, no decorative events, filter server-side.
• Write the tracking plan before the first line of code. Twenty well-named events beat 200 accidental ones in every tool here.
• Budget in rupees including GST, and re-check all three pricing pages every six months. Free tiers have moved in both directions, and a 20x cut has happened before.

Frequently asked questions

Is GA4 enough on its own for a small marketing team?

For acquisition reporting, usually yes. GA4 (Google Analytics 4) covers channels, campaigns, landing pages and conversions at no cost, and connects to Google Ads and Search Console natively. It becomes insufficient when you start asking person-level questions — which signup cohort retained, which first action predicts a repeat purchase — because GA4 is built around sessions rather than individual journeys over time. Its standard tier also keeps user- and event-level data for a maximum of 14 months, so enable the free BigQuery export early if you want long-run history.

What is the difference between product analytics and web analytics?

Web analytics measures sessions and traffic sources: how many people arrived, from where, and what it cost. Product analytics measures individual behaviour over time: retention curves, cohorts, and funnels you can re-slice after collection without re-tagging. GA4 is web analytics; Mixpanel and Amplitude are product analytics. Put plainly, web analytics tells you which ad to switch off, and product analytics tells you which onboarding step to rebuild. Most teams need both answers, so the usual lean-team setup is GA4 plus one product analytics tool sharing a single event naming convention.

Should a lean team pick Mixpanel or Amplitude?

Check the meter before the features, because the meter decides the bill. Mixpanel charges by events; Amplitude charges by monthly tracked users. High-traffic, lightly-engaged sites — content sites, D2C stores — usually stay free longer on the event meter, because they have many users firing few events each. Low-traffic, heavily-engaged products such as B2B SaaS tools usually stay free longer on the user meter, because each user fires hundreds of events. If your product or data team already runs one of them, use that one: a shared event taxonomy beats a marginally better tool in a second silo.

Do Mixpanel or Amplitude replace GA4?

No. Neither connects to Google Ads and Search Console the way GA4 does, so neither can do campaign and channel reporting as cheaply or as natively. Neither is free at the traffic volumes GA4 handles for nothing, either. The standard setup for a lean team is GA4 for acquisition plus one product analytics tool for behaviour, with one agreed event naming convention across both so the numbers can be compared. Removing GA4 saves you no money and costs you the common language you share with agencies and ad platforms.

How much do Mixpanel and Amplitude cost in India?

Both bill in US dollars, so the rupee cost moves with the exchange rate and normally attracts 18% GST as an import of services. At roughly ₹88 to the dollar, an entry paid plan around $25 a month works out near ₹2,600 including GST, and one around $49 a month lands near ₹5,100. Both scale upward with volume, so treat those as floors and model your own usage. For context, GA4 360 is commonly reported to start near $50,000 a year — roughly ₹44 lakh — which rules it out for most lean teams. Verify current figures on each vendor’s pricing page.

How many events should a lean team track to start?

Fifteen to twenty-five, each tied to a decision somebody actually makes. More events do not mean more insight; they mean faster consumption of your free allowance, a messier taxonomy, and movement toward GA4’s limit of 500 distinct event names. Collapse near-duplicates into one event with properties — a single “Product Viewed” event with a category property, rather than eight category-specific events. Skip decorative tracking such as scroll depth at four thresholds per page, which can account for a third of total volume while changing nothing you do.