Most of the people reading this blog have probably also seen demos or at least heard of Gemini and it’s magical powers of solving all BI problems known and unknown. (And you probably want to be looking at it. You can see if you get accepted to the CTP from the instructions here. I understand that you’ll automatically get the preview of Office 2010 if you’re accepted…)

From my perspective of implementing the PerformancePoint Planning Excel AddIn, I can say Gemini would have been a huge help. The compressed size of Gemini models is amazing compared to a local cube file. Not to mention the hassle of maintaining the cached copy outside the workbook and assorted issues of how to keep workbooks in sync with the data. Plus the query speeds really are fantastic. There’s just that little detail of what-if queries…

In the reporting and data analyst space, I think Gemini looks like a huge help. I see lots and lots of fear about how Gemini will be managed and the mess that ordinary users will make. (Disclaimer: I’m not on the Gemini team, I’m just talking about my interpretation of all the demo’s I’ve seen and the experimentation that I’ve done.) With what is in the CTP I would not be afraid of ordinary users. They probably won’t create complicated data models yet so not too much can go wrong. And for reporting to always have the report and the data in sync, seems like a huge help. For scenarios where the data needs to be refreshed at specific intervals or on demand the SharePoint integration I think hides most of the complexity from most of the users. I’ll come back to fear-mongering in another post.

For data modeling, I think the Gemini user experience is a huge leap forward as compared to BIDS. I think this will open up some data modeling capabilities to regular folks. But obviously it doesn’t do everything, but it feels like the kind of tool that end-users can work with. The kind of thing they were already doing manually in Excel with lots of temp tables and macros and manual intervention. For DBAs maybe it still looks limiting, but they are welcome to continue using BIDS. End users want to give up the fragile temp tables and macros.

From the PerformancePoint Planning experience, I know people want an easier way to define business rules. PEL was an interesting step, but it didn’t go far enough. It was easier than MDX, but not as easy as Excel formulas. Gemini takes it to the logical conclusion of as close to Excel formulas as possible. Now I have great faith that business users can achieve the result of defining business rules without a DBA intervening.

Clearly Gemini is not a magic cure-all. But it’s a huge step in the direction of making end-users able to self-serve their simple BI needs. When the data already exists but maybe in different silos. Or they need to provide some assumptions that can’t be edited in a LOB system. Self-service doesn’t have to mean you can prepare a 10 course meal. But you can pour the soda into a cup, and pull a salad together from the buffet line without having to bring in a chef from Le Cordon Blue too. Gemini is definitely the right direction, it will be interesting to see how close they get to the 10 course meal…

Again, I’m not on the Gemini team, and I’m not a spokesperson for Microsoft in any shape. I’m just talking through what I see, and why I am thinking differently than others about what Gemini can do for end-users. Flame away if you think I’m crazy, maybe I won’t delete the comments.