So, like everyone else this week I was impressed with the Google Wave demo, and like everyone else in the BI industry had some rudimentary thoughts about how it could be used in a BI context. Certainly a collaboration/discussion/information sharing tool like Wave is very relevant to BI: Microsoft is of course heavily promoting Sharepoint for BI (although I don’t see it used all that much at my customers, and indeed many BI consultants don’t like using it because it adds a lot of extra complexity) and cloud-based BI tools like Good Data are already doing something similar. What it could be used for is one thing; whether it will actually gain any BI functionality is another and that’s why I was interested to see the folks at DSPanel not only blog about the BI applications of Wave:
…but also announce that their Performance Canvas product will support it:
It turns out that the Wave API (this article has a good discussion of it) makes it very easy for them to do this. A lot of people are talking about Wave as a Sharepoint-killer, and while I’m not sure that’s a fair comparison I think it’s significant that DSPanel, a company that has a strong history in Sharepoint and Microsoft BI, is making this move. It’s not only an intelligent, positive step for them, but I can’t help but wonder whether Microsoft’s encroachment onto DSPanel’s old market with PerformancePoint has helped spur them on. It’s reminiscent of how Panorama started looking towards SAP and Google after the Proclarity acquisition put them in direct competition with Microsoft…
Meanwhile, Google Squared has also gone live and I had a play with it yesterday (see here for a quick overview). I wasn’t particularly impressed with the quality of the data I was getting back in my squares though. Take the following search:
The first results displayed are very good, but then click Add Next Ten Items and take a look at the description for the TopCount function, or the picture for the VarianceP function:
That said, it’s still early days and of course it does a much better job with this search than Wolfram Alpha, which has no idea what MDX is and won’t until someone deliberately loads that data into it. I guess tools like Google Squared will return better data the closer we get to a semantic web.
I suppose what I (and everyone else) like about both of these tools is that they are different, they represent a new take on a problem, unencumbered by the past. With regard to Wave, a lot of people have been pointing out how Microsoft could not come up with something similar because they are weighed down by their investment in existing enterprise software and the existing way of doing things; the need to keep existing customers of Exchange, Office, Live Messenger etc happy by doing more of the same thing, adding more features, means they can’t take a step back and do something radically new. Take the example of how, after overwhelming pressure from existing SQL Server users, SQL Data Services has basically become a cloud-based, hosted version of SQL Server with all the limitations that kind of fudge involves. I’m sure cloud-based databases will one day be able to do all of the kind of things we can do today with databases, but I very much doubt they will look like today’s databases just running on the cloud. It seems like a failure of imagination and of nerve on the part of Microsoft.
It follows from what I’ve just said that while I would like to see some kind of cloud-based Analysis Services one day, I would be more excited by some radically new form of cloud-based database for BI. With all the emphasis today on collaboration and doing BI in Excel (as with Gemini), I can’t help but think that I’d like to see some kind of hybrid of OLAP and spreadsheets – after all, in the past they were much more closely interlinked. When I saw the demos of Fluidinfo on Robert Scoble’s blog I had a sense of this being something like what I’d want, with the emphasis more on spreadsheet than Wiki; similarly when I see what eXpresso is doing with Excel collaboration it also seems to be another part of the solution; and there are any number of other tools out that I could mention that do OLAP-y, spreadsheet-y type stuff (Gemini again, for example) that are almost there but somehow don’t fuse the database and spreadsheet as tightly as I’d like. Probably the closest I’ve seen anyone come to what I’ve got in mind is Richard Tanler in this article:
But even then he makes a distinction between the spreadsheet and the data warehouse. I’d like to see, instead of an Analysis Services cube, a kind of cloud-based mega-spreadsheet, parts of which I could structure in a cube-like way, that I could load data into, where only I could modify the cube-like structures containing the data, where I could define multi-dimensional queries and calculations in an MDX-y but also Excel-y and perhaps SQL-y type way – where a range or a worksheet also behaved like a table, and where multiple ranges or worksheets could be joined, where they could be stacked together into multidimensional structures, where they could even be made to represent objects. It would also be important that my users worked in essentially the same environment, accessing this data in what would in effect be their own part of the spreadsheet, entering their own data into other parts of it, and doing the things they love to do in Excel today with data either through formulas, tables bound to queries, pivot tables or charts. The spreadsheet database would of course be integrated into the rest of the online environment so users could take that data, share it, comment on it and collaborate using something like Wave; and also so that I as a developer could suck in data in from other cloud-based data stores and other places on the (semantic) web – for example being able to bind a Google Square into a range in a worksheet.
Ah well, enough dreaming. I’m glad I’ve got that off my chest: some of those ideas have been floating around my head for a few months now. Time to get on with some real work!