What is our primary use case?
My main use case for
Amplitude, which relates to my main job role, is analyzing data. That involves creating dashboards, connecting click-level data sources to
Amplitude, and then consuming reports and creating analyses for stakeholder-facing reports, which would be consumed by company leadership, the analytics team, product managers, and others.
A specific example of a dashboard we built with Amplitude is a click-level report that I created recently when I was with TurboTax. We made it to study the clicks on every page. Essentially, what we visualized was a Sankey chart of how people are moving through pages. Since we changed the entire product stack, it was important for us to verify where the drop-offs were. Eventually, we moved the idea of a Sankey chart to a drop-off chart where we studied the bounce rate for each page, and we had a bar graph showing the number of people on one page and throughout their journey, where people dropped off. That drop-off report was a big project, and I think that helped the team recognize what the bottlenecks were and what the problem areas were in the product.
The reaction from our team and stakeholders to the findings from the drop-off report was that it had already been discussed with the major stakeholder product managers, so it was a requirement we mutually discussed and came to. Since Amplitude has incredibly easy functionality for accessing the past year of data, it became a tool that even the PMs could play around with, allowing us to leave them to self-serve. That acted as a big positive for us in the analytics team because the more our stakeholders use self-serve reports, the less work it is on our shoulders. That was especially great from the PMs because we could reconcile our Amplitude data with other data sources, such as the Databricks data lake, in a short span of time. Changes were made right from UX level changes on the pages to broader decisions on whether we need to clamp down on the entire tax filing experience.
What is most valuable?
I think the best feature Amplitude offers is its data visualization capabilities. As a subject matter expert who has been working on it for seven to eight years, I consider Amplitude to be the closest any application software gets, and it just makes life easy for developers as well as people with less developer experience. Even for product managers who do not have data visualization development experience, they can navigate very easily, and so is the case for any other stakeholder. That is the best feature.
What stands out about Amplitude's data visualization compared to other tools I have used is the easy integration with clickstream data. It was seamless for that data to be loaded into Amplitude, with a good volume that easily loads onto Amplitude without having to create intermediate tables, which was the case with Tableau. Everything on top of that is easy usage and functionality because we used to do it on Chrome without needing a desktop server to create data visualizations. This entire package is really robust.
Collaboration was a significant part. What improves collaboration is the self-serve functionality, which was a big deal for PMs to have access to just that data and also the base layer of how that data is structured, which connects to clicks that every report refers to. Business stakeholders can pull different cuts and slices without relying on analytics, which is usually quite loaded with tasks. That was the best thing, as collaboration allowed business stakeholders to attempt to go through the reports and create their own reports, which was beneficial.
I notice faster decision-making for sure, as business stakeholders could themselves get their own reports.
What needs improvement?
I think areas for Amplitude's improvement include data retention. The maximum data I have worked with, the clickstream data on Amplitude, is two years, whereas we rely on Tableau for historical data of at least two years. I am not sure whether it was my organization paying for lower retention or not, but having historical data would take away the reliance on Tableau. Secondly, reconciling clickstream data with
Databricks or other
AWS systems could help analysts spend less time verifying the accuracy of both sources, which would be really helpful. Lastly, a more focused approach on error logs is necessary. I have found that, at least with Intuit, we used Splunk more for error logs than Amplitude. Although Amplitude had the capability, developers seemed to prefer Splunk, so if that could be improved, it would be good.
For how long have I used the solution?
I have been working in my current field for about seven and a half to eight years. I have been using Amplitude for five to six years.
What do I think about the stability of the solution?
I did not notice any delays or issues with Amplitude's performance and speed when handling large datasets. One of the strongest points of Amplitude is its functionality for making visualizations, and speed while handling large datasets is really good. There has never been an issue where it hung or took a lot of time, and that is where Amplitude has the edge in handling large datasets with clickstream data.
What other advice do I have?
I did not work a lot on the integration side of Amplitude, so the data engineers worked on integrating clickstream data, and my problems revolved around reconciling this clickstream data on Amplitude with the Databricks data. It made for a tiresome experience because essentially they should be the same, but they were not due to different reasons on the data engineering side. I did not really integrate Amplitude with other tools per se.
Amplitude's security and access controls are completely in-house, and organization-wide accesses were provided right when we onboarded onto our analytics roles, so that was never a problem. Access controls were also straightforward, and we did not have any issues there.
Managing data privacy and compliance requirements with Amplitude was easy. As far as I remember, the challenges with data privacy and compliance requirements did not raise any issues flagged at any point by the data engineers or the Amplitude admin. We could seamlessly use the data from an analyst perspective.
We did not use the mobile app, so I cannot comment more about it.
I would describe Amplitude's reporting and analytics capabilities compared to other tools I have used as really good. I would say it is on par or just shy of Tableau, which, for me, is the best tool in the market, so it is right up there.
I went through the training resources and documentation when I joined in 2019, and as far as I remember, they were useful. The training may have included videos or articles, but I do not fully recall. The prompts on the website and the nudges were really helpful.
My advice for others looking into using Amplitude is that if it is clickstream data, Amplitude is likely the best option. Having accessible data which can be easily customized and manipulated by even business stakeholders with less technical knowledge is one of the best things. I rate Amplitude an eight because the changes I mentioned, including longer data retention, easier integration with AWS data sources, and a better focus on error logs, would make it a ten for me.