The question of Mixpanel vs Google Analytics pop-ups in my inbox on a weekly basis. Companies of all sizes are confused as to how to use these tools. Do they play well together? Do they overlap in functionality? One is free (for the most part) and the other can get pricey quickly. What do I recommend?
This is the article for you. I’m not going to cover every single thing about each tool but instead, look at the use cases for each tool. I’ll also share my ideal recommended setup and how you could hack GA to do everything if you really wanted to.
Let’s jump right in by looking at who Mixpanel and GA is really for.
Let’s start by looking at what situations and companies can benefit the most from Mixpanel. Mixpanel is primarily built for product teams who are trying to understand the behavior of their users. This includes software companies, ecommerce, and media companies. If you have a digital product and you need to look at how users are interacting with it, Mixpanel is for you.
Mixpanel is built exclusively on event tracking which is a flexible way to structure your data. From there, you can run reports like funnels which let you see the drop off in between steps.
You could also run retention reports to analyze how sticky your product is over time.
All of these reports are built for complex products. If you have a “simple” website (relatively speaking) with few user interactions, then Mixpanel might be overkill. Ecommerce companies are interesting because they could benefit from a tool like Mixpanel but they could also do just as fine with Google Analytics. Software products are perfect because they are intrinsically complex in how they are used.
If you find yourself using Mixpanel to just track pageviews, you’re paying too much. If you’re planning on designing a data schema that is rich and complex, then this could be a great fit.
Google Analytics (GA) is still the gold standard for marketing attribution in 2020. Even at the free pricing plan, it is a robust tool that can do a lot. Google Analytics has been built on a pageview model and captures most of the data that you care about out of the box. You can of course, track events with GA which I’ll cover in the next section.
In short, GA is fantastic for understanding where users are coming from (marketing attribution) and can be fantastic for “simple” websites that have relatively low user interactions. Informational websites, blogs, and ecommerce sites all fall under here. Ecommerce, in particular, has the option of setting up Ecommerce Tracking which would add the ability to track revenue.
Google has been working hard to add new reports like cohort analysis, better funnels, and the ability to track users by ID (instead of anonymously). This leads to companies to think: maybe GA can do everything I need for my digital product. I still believe that this is the wrong approach as you would be working with a much more limited product than what Mixpanel can offer.
However, if you want to hack GA to its full value, you could get the answers that you’re looking for. You would need to use EVERYTHING GA offers. I mean everything. Custom dimensions, user ID, ecommerce tracking, event tracking. You would then need to export this data into a spreadsheet to tweak it and visualize it. It’s possible but somewhat inefficient.
The main question that comes up with GA is around events. Ah events. Why can’t we use events in GA to replicate Mixpanel? I’ll tell you why.
GA does have events. That is true. I actually had someone make a passionate argument about why I was lying about GA’s event tracking capabilities.
The event tracking in GA is okay. That’s how I would describe it in one sentence. For starters, you can only track 4 properties in total: category, action, label, and value (this is a number). Mixpanel lets you track up to 2000.
You likely don’t need two thousand event properties but you do need more than 4. GA forces you to compromise and track more events which makes the data harder to work with.
The event analysis in GA is limited too. For example, if I wanted to slice an event within the GA interface, I could add a secondary dimension but only one. In Mixpanel, I could easily add 20 or 30 segmentations.
The funnels in GA (likely the biggest weakness of the free version) requires set up beforehand and they still aren’t that good. Mixpanel lets you build funnels on the fly.
I know what you’re thinking. Just simply export the data and analyze it in a spreadsheet. This works fine but would you be better off with something that is more robust? I would think so.
As I said above, if your product is relatively simple OR if you wanted to hack GA to its maximum potential, these limitations would be fine. If you’re dealing with something more complex, Mixpanel is the way to go.
The new Web + Mobile view in GA is meant to solve the cross-platform problem for companies that have web and mobile apps. This is still in beta and it’s ok as well. It combines web activity with mobile activity (from Firebase) into one.
The reason why this is ok (as of today) is that it doesn’t fully have all the features of web views and it still has limitations around events. I could see this getting better in the future but it will take time.
Keep in mind that Google Analytics today is fundamentally very similar to the original Urchin product that Google bought. Look at screenshots and you will know what I mean. Google is improving this product but far slower compared to companies like Mixpanel.
If you have simple digital products on the web and mobile, then this view can be fine.
That being said, I don’t think GA and Mixpanel are foes. I think they are closer to friends and perhaps frenemies. For my clients, the ideal setup is to have both. Yes, both at the same time.
In reality, I think companies should use both tools. Some people struggle with this idea because they think it’s inefficient. Why have 2 tools when you can have one? The answer is that each is made for something different. This is also how you build modular marketing stacks.
Google Analytics marketing attribution is fantastic because they basically invented the concept of UTM parameters. You can easily see different attribution models and you even get access to data-driven models.
Mixpanel user behavior analysis is top-notch and you also get access to reports powered by machine learning. Add messaging and A/B testing (mobile only) and you get a robust analysis platform..
I see GA taking care of anything that happens before the user signs up and I see Mixpanel taking care of everything that happens after they sign up. GA hands the baton to Mixpanel like a relay race. The baton is the metaphorical user.
Advanced teams are then able to export the data from both tools into a data warehouse and analyze it together. If you’re not here yet, don’t worry. This is an advanced strategy and a result of a data-driven culture.
There’s only one more thing to talk about. How much do these tools cost?
Pricing matters especially if you’re talking about millions of users and events. Google Analytics pricing is simple. Free forever or $150,000+/year for GA 360. GA 360 has other things like better funnels and better attribution.
The free forever is true and only has one catch. At a certain point, GA will start to sample the data you see in reports. You can tell by looking for this sign:
The raw data isn’t sampled and you could always export it via the API which is what most high volume sites have to do. At this point though, you lose all benefits of the GA interface and you’re stuck with spreadsheets and BI tools.
Mixpanel on the other hand charges for MTUs or Monthly Tracked Users. This includes anonymous users and known users with accounts. Pricing starts at $89/month for 1,000 users and there’s also a free plan for beta testing.
The paid public pricing for Mixpanel goes up to 25,000 MTUs and then you’re dealing with sales reps. High volume products can quickly ramp up their Mixpanel bill especially consumer products with a high number of anonymous users.
This means that most companies will pay $0 for GA and $XX for Mixpanel depending on users. Companies that pay for GA 360 also typically commit to the Google world and move away from tools like Mixpanel though these are larger companies.
The answer to the question of Mixpanel vs Google Analytics is: it depends. This is the classic fallback answer of experts but it’s the truth. My recommendation is to use both for what they do best. However, you could hack GA to do a lot of what Mixpanel does but expect lots of manual work.