This past Friday we had a board meeting and I demonstrated Tigger (project name for our B.I. product). The response was fantastic. One board member asked me to set up a webinar with him and the main B.I. person at his company to demonstrate it as they may want to OEM it. A couple of them started drilling us on do we have people and processes in place to grow to XXX million next year as this should be gigantic.
But the nicest comment of all was (going from memory here) "other companies have problems getting software shipped. Windward has a problem that it delivers products and new versions so quickly marketing & sales are scrambling to catch up." That made me feel really good about our development team – and yes they are outstanding.
Ok, so on to our status. We now have ad-hoc reporting complete for SQL datasources hitting it from Microsoft Word. Excel and PowerPoint are going through final testing so they should also be complete by the end of this week (the parts specific to Word/Excel/PowerPoint are pretty minor outside of drill-down).
So what does it look like? As you can see in the picture you are in Word, with a task pane at the top (it can be moved elsewhere) with all the ad-hoc variables. Off each ad-hoc variable you can drop down a menu to sort, filter, and/or select from a list. If you have more than 10,000 items, the list only shows the first 10,000. This all operates very similar to Excel.
There is one difference from Excel. When you type a search term in for the list, if the full list is over 10,000 items, then it will do a SQL select using the search text to re-access the database to pull the first 10,000 that match the search term. By doing this we avoid having to pull the entire list into memory as that could be hundreds of millions of rows (and that would not be good).
And filters are also pretty straightforward. One difference with Excel is Excel filters by a column. We filter by a select variable and that variable is the column(s) returned in a row of data by the select. So it will show all possible columns, then the operations allowed for that type of data, and then the value. The right-most edit box is cool in that if the column type is a date, the right-most editor is a date picker. If a number, it is a number spinner, etc.
We're also thinking of placing a slider below the control for both dates and numbers – what do you think? And maybe a MRU list if it's a text editor?
The next step is alpha testing by end users for usability. We have a couple of our customers that will be trying it out this week and next. But if you're a sophisticated B.I. user and would like to give us feedback, please email us at firstname.lastname@example.org and we'll try to set you up with an alpha version. I'm confident we've made this very easy to use. But that doesn't mean it's easy to understand what to do for all use cases – our alpha testing will be the real test of that.
I originally thought we would have XPath (XML datasources) support next, followed by drill-down. But the XPath group is finding it harder than expected while the drill-down group is making much faster progress than expected (while cursing the Word team for not having a url click event). So we may have drill-down on SQL datasources out as an alpha in 2 – 3 weeks. XPath support will probably be a month or so behind that.
The Magic Behind the Curtain
I've been asked by a bunch of people what is required to configure this system. I'll try to write that up shortly in a new blog entry. But the short answer is very little. No need to create cubes. No need for a DBA or developer to design the templates. Just two very simple straightforward selects. Pretty much what people designing templates for our reporting and document generation systems do. And we have tens of thousands of non DBA/developers successfully using that.
So I'm pretty comfortable saying you can install this and ½ hour later a business user can be running their ad-hoc queries.