Overall, my experience with Amazon Quicksight has been amazing. I wasn't the person responsible for the initial integration/implementation within our software ecosystem, but I am one of the primary dashboard creators on our team and I'm definitely not a developer. As a semi-technical user I've found that the platform is very robust and has tons of features that make Business Intelligence reporting easy to do. Although I still haven't gotten the hang of ALL the advance features, I do feel like I'm slowly catching on to them and am probably in the 80th percentile of users. Now that I've had this tool, I definitely couldn't live without it.

Ease of use with the reporting mechanisms is second to none. The ability to incorporate Amazon insights (AI) is amazing. I've also noticed that they release tons of new features constantly. Dashboards look really good as well.

Pretty difficult to configure initially. If you're looking to leverage all the tools within your current system, then you'll really want to have someone certified in AWS or at least a team of devs with advanced skills. We actually reconfigured our implementation within our software after about 6 months of use and it made our lives much better.

Overall, our team has been very please with what we get with Amazon QuickSight. The support team and our account reps have assisted us every step of the way. Their unique (from our previous solution) pricing structure took a bit of time to quote out, but once the math was done it was clear which solution was best. While we didn't keep as many dashboard author seats as first expected, the time and workload associated with onboarding and implementing was not something that overly taxed our engineers or analysts.

The pricing structure alone got us interested, but the ease of setup and the user experience kept our attention. On the data-prep side, linking our RDS and Aurora instances was as simple as a login prompt. Mapping and prepping data was just as simple too. The process of building out our first interactive dashboards and then embedding them for our users occurred in 1/4 of the time as our previous BI solution.

We still find a desktop SQL client the easiest route for the initial dataset queries. The editor in QuickSight is very beginner-friendly though. Someone with more time writing queries will likely only copy/paste into the editor, however.

What I like the most about Quicksight is ease of use - even if you're not a BI, it's fairly simple to create dashboards and real-time updates for visualizing data. Learning curve to actively creating these is way faster than similar tools on the market. If you have more developed knowledge - no worries, integration with AWS services can keep you busy to create something a bit more comprehensive. Being part of AWS galaxy, it's quite 'plug and play' if you already use other AWS services.

Could do with some improvement on filtering (as each tab is handled separately, you can't 'filter all' on your reports to slice data on different reports). Also borrowing some more customization and pivoting option as other actors in the market do (see extreme flexibility from end-customer side that Looker has) would be a massive leg up, as current system is a bit 'rigid' as a end user - ie., very little opportunity to customize.

Ultimately, I'm not enamoured with QuickSight at all. Yes, it is an easy-to-use, highly accessible, AWS-compatible BI product but it feels half-baked with the rest of AWS, and it has some curious approaches, some complexities that are unintuitive, and some other things that make it just that bit more frustrating than it needs to be.

QuickSight is simple to set up - it's a public cloud AWS service so if you're already using AWS it's already there and you merely sign in to your AWS console and navigate over. It's also easy to set up your first visuals; you create a data source, create an analysis, and away you go. It helps you make interesting and effective visuals too, giving recommendations based on the nature of the data too. So, the barrier to entry for QuickSight and to start using it is extremely low. There's nothing to it.

It's clear QuickSight is a product AWS acquired rather than natively developed because so many things are different or inconsistent with the usual AWS standards - menus, security groups, user accounts, etc. Some things you could only do via the AWS CLI. As an example, until only just recently, QuickSight had no GUI support for groups. We could make groups via the CLI, and then assign data sources and analysis to the group, but it was not available or visible in the GUI in any way. Thankfully, this feature has been added during March or April 2022. Another issue is that QuickSight sometimes doesn't work as you expect. We encountered issues with pivot tables whereby we got different results when pivoting/grouping on a specific column than when we exported the exact same data into Excel and pivoted there. We manually compared each individual row and confirmed we were looking at the exact same data. The Excel pivot was correct, while our QuickSight pivot was doing some averaging. We logged this with AWS support for help and while it was resolved the issue is that some things that ought to be simple and intuitive become a bit more complex in QuickSight. Also, QuickSight can store results for quicker access in storage it calls SPICE, and you buy SPICE quota. However, if you run out of SPICE it won't run your scheduled reports and they simply fail.

Overall Amazon Quicksight has been a great tool!

It has helped us display different metrics and improve our performance more than before!

Some of the GUI is not the best but it's just the process of customizing it more.