I'm very grateful to be able to use it, and I have a master's degree with a focus in Geospatial Analysis. There can be a bit of a learning curve, and I try to build user-friendly ways for volunteers to see & collect data. Meanwhile, if a colleague is less confident with building such a system, it may be more difficult for them to implement.
I have not seen a better mapping tool than CartoDB. You get the familiarity of Google Maps with arbitrarily complex geographic data visualization on top. CartoDB excels at large data sets where Google Maps API completely chokes when attempting to handle more than ~1000 data points. I was able to plot 500,000 points on a map with reasonable speed and able to perform complex aggregations to display boundaries of areas containing certain types of data, intersections of those sections, and more.
Learning curve - CartoDB might be difficult to use if you don't have a bit of SQL or data structures background. If you're not familiar with floats, strings, etc., you might upload an Excel file and be confused about how to manipulate it to get the software to create the maps that you want.
Performance - When I used it, there were some occasional issues with loading and parsing large data files.
Simply because the program deserves it. It seems to me that it is a fundamental tool for the storage, analysis, and interpretation of medium and large-scale phenomena, unmanageable with traditional engineering software. Its versatility in the handling of the different "layers" with which the data is handled and interpolation tools, make this software a powerful ally both for companies and for the educational part of the universities.
Once set up, the tools are extremely easy to use. I had a staff member develop a tool for field data collection, that included an external and internal dashboards to monitor progress in days. The field workers that collected the data, barely knew how to use a computer, and within minutes they could use the application that was configured for them.
ArcGIS' web support is quite good. They also have a broad user base that is active in answering community members' questions. For more sophisticated questions, ArcGIS technical support is good at answering questions, although answers usually are not instantaneously available. I'd also say that ArcGIS is working hard at making its technical support more reachable, at least it appears that way.
My students love the "drop" feature in Google Maps, but besides that it truly doesn't compare. I love that you can add, delete, or change layers to this map to better understand its larger affect. There are many more ways to manipulate maps on ArcGIS than on Google Maps. I can also add personal details and information if I want to create a specific map, something that I am unable to do with Google
Python is definitely a more powerful tool for data munging and analysis, but the python packages for geo-related data viz (bokeh, matplotlib, seaborn) are cumbersome to use. I would recommend doing your data analysis in Python and then exporting the final data to CartoDB for visualization. One benefit of doing this is that CartoDB can automatically publish your viz to a link or object, so you don't have to export it and host it yourself. Another benefit is that CartoDB automatically updates the viz once you change the data, eliminating the need to continuously regenerate image files. I haven't used Tableau too extensively, but from the experience I've had with it -- Tableau is better suited for traditional analytical visualization (charts, graphs, etc.) than for geospatial mapping and visualization.
