There’s something about a showdown that is exciting. We love to compare things. Apple vs Google, Coke vs Pepsi, and even things like summer vs winter. The same applies to software tools. In this article, we’ll do a showdown of Google Analytics vs Amplitude.
One tool has been around for a long time and has defined fundamental concepts for web analytics. The other tool is relatively newer but has quickly made a name for itself. If you have seen my Mixpanel vs Google Analytics comparison, then you know what to expect.
To begin with, we need to go back in time to understand the present. Sidenote, this is the concept behind one of my favorite podcasts (Throughline by NPR). Anyway, back to our article.
Let’s look at the original intent for each tool and how that affects how we use it today.
I believe the original design of a tool can help us understand how to use it. Yes, you might find interesting ways to hack a tool but that may not be efficient. I could try and hack my bike to work on water but I would be better off with a boat.
Google Analytics (GA) was originally designed for content websites in the early internet days (the late 90s). It evolved into being a great tool for understanding WHERE users are coming. Basically, marketing performance. Google Analytics is responsible for the creation of UTM parameters and most tools have copied how GA captures and stores them.
Based on this, you can imagine that anything does with marketing attribution is quite strong in GA. You can easily see last & first touch and even run multi-touch attribution reports. You can also see other performance like landing page conversion rate, time on site, etc. All the kind of things a marketer would need to understand their website.
GA also designed their data around the idea of anonymous users which is what most marketing traffic is. Most of the reports in GA are suited around looking at users in aggregate. They have added things like userID which lets you track individual users but again, it’s an addition not an original design. Events in GA are also limited to 3-4 event properties which is not enough for digital products.
Eventually, GA added Firebase which is their attempt to port this work for mobile apps. They kept the same principles but added mobile-specific metrics. Firebase is actually more similar to Amplitude in functionality but it still feels like a GA clone made for mobile marketing.
Amplitude was designed for analyzing digital products like software apps, ecommerce, and more. It was born at a time when apps (web or mobile) were already quite popular and there was a need for deeper analysis. The importance of concepts like the North Star metric has made Amplitude even more relevant.
Amplitude is built on a user model and they assume that you will be tracking users by their name, email, or something else that is unique. It’s not just about aggregate numbers but looking at what John Smith did in detail.
This user model can be quite powerful but it only works for situations where you’re mostly tracking actual customers or users. They also expanded events to make it extremely flexible. Events let you track anything you want and you could understand almost any kind of product using them.
Finally, Amplitude started with several reports better suited for digital products. Flexible funnels and cohort analyses were there from the start. The event model makes these reports quite powerful as well. Amplitude also requires a significant shifting of your culture (data-driven or not).
In reality, Google Analytics and Amplitude are more complementary than you might imagine. I typically get clients to use GA to analyze everything that happens before a user signs up and Amplitude for analyzing anything that happens after the signup. They also form the foundation for a marketing automation stack.
We could also use Amplitude for attribution but it isn’t as good as Google Analytics. Instead, the hybrid models work much better. That being said, let’s look at how their implementation differs and best practices that you should know.
Your team could get GA up and running in a few minutes but a tool like Amplitude will require 15 – 20 hours of solid work to get it going. This work is worthwhile especially if you have a digital product as mentioned in the previous section.
Mobile is a special case. Amplitude also offers client-side libraries here while GA does the same but it does it through Firebase. You can connect your Firebase data with your Google Analytics web data but this is still a bit wonky as of mid-2020.
Once you get both tools up and running, you’ll need to build reports. Let’s look at whether this will be enjoyable or a nightmare.
Ah, building reports. This is where we realize how good (or bad) our data is.
For Google Analytics, most of the reports that you care about will already be built. In fact, you have so much data out of the box that you actually need to find a way to simplify this for your own sanity.
You could explore a pre-built dashboard that summarizes the most important KPIs in a single place. You could also build custom reports that do the same. Either way, you’re remixing stuff that is already built and you will be up and running almost right away.
Amplitude doesn’t come with any pre-existing reports except things like Sessions. Building reports in Amplitude is quite easy though once you learn the tool. The query interface that you will be using is a “visual SQL”.
The key to Amplitude reports is your data. If you have all the events that you need and these events include all the relevant event properties, then it’s just a matter of building your query. If you don’t have this or the data doesn’t make sense, it will feel impossible.
Before we jump into pricing, let’s take a second to look at user tracking in both of these tools. Companies looking to hack GA to do similar things to Amplitude will actually find this relevant.
As mentioned above, Amplitude is built around the idea of a user model. John Smith is taking different actions (events) and we can track them all back to him/her. Google Analytics could in theory tracking users by an ID as well, and we call this userID.
First, some limitations. Google Analytics doesn’t let you send any PII like email as the userID. This may not be a big deal but some companies only have access to the email. Amplitude supports any kind of ID but they recommend that you use something persistent like a database ID.
Second, GA has a few reports where you can see individual users and their actions but it’s a little tricky to actually use this data in other reports. Saving the userID as a custom dimension helps.
Besides these two points, it basically works in a similar way. I think you get more use of the Amplitude user tracking but that’s a function of the reports and how how they track users. I still believe userID in GA feels hacked on but if that’s your only option, then it will do just fine.
Finally, let’s look at pricing. Surprisingly, both of these tools have similar pricing models.
It’s rare that we see two tools that are so similar in pricing. Both of them have strong free plans. Google Analytics limits your to 10M hits a month, where a hit is a pageview, event, etc. This is a soft limit as I have seen many accounts go beyond this with minimal issues. While your account won’t be shut down, you’ll start to see sampled data inside GA.
Amplitude limits you to 10M events a month and unlimited users. This can be quite a bit of B2B products but almost nothing for consumer products. This is a hard limit that Amplitude will enforce so you’ll need to do some thinking around your tracking design to start within it if that’s your goal.
The first paid level of Google Analytics is GA 360 which starts around $150,000/year. Pretty steep upgrade right? Upgrading here usually also means that you’re committing to the 360 world of Google products.
If you’re over the 10M hits limits from GA, you should be fine in theory but you never quite know. Is 20M fine? What about 50M? What about 100M? At some point, you’re bound to receive a call from Google asking you to upgrade. If this uncertainty is too much, you may need to explore the 360 product.
Amplitude’s growth plan starts at $40,000/year but that depends on multiple factors. For smaller companies, you would be better limiting what you capture to start within the free plan since most of the advanced reports in paid plans are better suited for larger companies.
Most of my clients get GA for free but pay for Amplitude. This is the likely scenario if you use both tools (which you should).
As you can see, there are similarities and differences between these two tools. They are more powerful when used together but they also require vastly different capabilities. I hope this article helped clear up any confusion you had about them. Be sure to also check out my full guide on how to use Amplitude and all its data.