April 17, 2012 at 9:50 pm
Comments posted to this topic are about the item Social Data Analysis
Follow me on Twitter: http://www.twitter.com/way0utwest
Forum Etiquette: How to post data/code on a forum to get the best help
My Blog: www.voiceofthedba.com
April 17, 2012 at 11:06 pm
Since this is a repost from an editorial written in 2007 - 5 years ago - what is your current viewpoint?
A lot of things surely have changed: a shift to more self-service BI with Report Builder, PowerPivot and PowerView, in-memory processing with xVelocity (tabular models and columnstore indexes) and so on.
Need an answer? No, you need a question
My blog at https://sqlkover.com.
MCSE Business Intelligence - Microsoft Data Platform MVP
April 18, 2012 at 7:37 am
Frankly - so long as there is as Excel in your company you have a BI application. Understanding that there is always BI/MI running around the place will help you realise that having at least one centralised source which things can flow from will stop you from waking up in a cold sweat at night because tomorrow is the day when board reports get compiled and sales and marketing will produce competing versions of profits and sales figures and you get the finger pointed at you, lowly IT member, who had nothing to do with their manual data entry and on-the-fly spreadsheets.
Code your own, buy one, download an open source version, keep it all in excel - so long as users can consume it in excel and you can determine what goes into excel everyone will be fairly happy. It's then the users who want animated charts and sparkly things you have to decide whether to accomodate 😀
You don't even need to spend any money as SSRS can be gotten for free under the Express edition. You get out of the box decent functionality that you can give to the users in the form of report builder so that they can all pull from the datasets in a fairly excel like environment. Give them PowerPivot too. I think Excel integration is the single biggest factor for me when it comes to reporting suites and it's a rare company indeed where users don't try to horde spreadsheets!
April 18, 2012 at 8:29 am
That was great, before I saw the "this is a repost" I thought you were losing your mind. I am always so appreciative of the knowledge and insight you share and I think the retrospective ability to expose your prior opinion was a humble and enlightening one.
I think the Social Data Analysis reference is particularly applicable to the current times as folks like Google, LinkedIn and Facebook spin analytics into their offerings and not always with the high-dollar unreachable toolsets.
To me BI is a philosophical solution that has been around longer than software and is done everywhere and has been for centuries; a farmer makes a decision on what crop to grow based on demand, a rancher looks at his land and based on his knowledge knows how many head of cattle he can winter …
I have been a BI professional for about 15 years and I believe Excel was really one of the first widely accepted BI Software Tools and so many of the newer tools try poorly to emulate it. So in the past I have always been disappointed in the so-so showing of the MS BI Stack but, I really believe that MS has squashed this challenge with SQL Server 2012. With the functionality and capabilities in a simplistic licensing model (and yes compared to the other vendors Server, Core, CAL is simple) the right people will be able to stand-up a reasonable priced BI Solution capable of meeting small to medium demands and potentially even larger demands with more investment. To clarify I don’t believe the MS Stack solution is better than the high dollar solutions, but is comparable in many facets including some of the key ones. There are some key limitations still, but hopefully with the right uptake these will get resolved as well soon (like making SSIS a repository based ETL tool capable of team development.)
I have always been excited about BI, but now I am truly excited about MS BI as it is showing the right indicators.
April 18, 2012 at 10:09 am
Koen Verbeeck (4/17/2012)
Since this is a repost from an editorial written in 2007 - 5 years ago - what is your current viewpoint?A lot of things surely have changed: a shift to more self-service BI with Report Builder, PowerPivot and PowerView, in-memory processing with xVelocity (tabular models and columnstore indexes) and so on.
Absolutely.
Over the last two years I have seen requests for new reports drop from one a week to zero while requests for asccess to the cubes are ever increasing.
We will be taking down most of our (internal) BI site pages; we will keep supergraphics that are impossible (or impractical) to do in Excel.
April 18, 2012 at 10:21 am
stephanie.sullivan (4/18/2012)
Frankly - so long as there is as Excel in your company you have a BI application. Understanding that there is always BI/MI running around the place will help you realise that having at least one centralised source which things can flow from will stop you from waking up in a cold sweat at night because tomorrow is the day when board reports get compiled and sales and marketing will produce competing versions of profits and sales figures and you get the finger pointed at you, lowly IT member, who had nothing to do with their manual data entry and on-the-fly spreadsheets.Code your own, buy one, download an open source version, keep it all in excel - so long as users can consume it in excel and you can determine what goes into excel everyone will be fairly happy. It's then the users who want animated charts and sparkly things you have to decide whether to accomodate 😀
You don't even need to spend any money as SSRS can be gotten for free under the Express edition. You get out of the box decent functionality that you can give to the users in the form of report builder so that they can all pull from the datasets in a fairly excel like environment. Give them PowerPivot too. I think Excel integration is the single biggest factor for me when it comes to reporting suites and it's a rare company indeed where users don't try to horde spreadsheets!
Not quite.
BI and OI have five steps:
- gather data
- cleanse data
- consolidate data
- make it accessible (practical)
- present data
- keep the data as the one and only truth about the past of the company
Excel addresses only the fourth point, and it works best if it accesses well-designed cubes. I cannot imagine having say fifty users, each of them running complex queries against data mart tables that have 20+B rows. (Which we have.)
"The only truth" is the biggest problem. As long as you have Excel workbooks with embedded pivot tables, accuracy of the data will depend on its freshness, and goes down the drain entirely if even a single business entity definition changes. Spreadsheet owners are usually unaware of the change.
April 19, 2012 at 2:26 am
Not quite.
BI and OI have five steps:
- gather data
- cleanse data
- consolidate data
- make it accessible (practical)
- present data
- keep the data as the one and only truth about the past of the company
Excel addresses only the fourth point, and it works best if it accesses well-designed cubes. I cannot imagine having say fifty users, each of them running complex queries against data mart tables that have 20+B rows. (Which we have.)
"The only truth" is the biggest problem. As long as you have Excel workbooks with embedded pivot tables, accuracy of the data will depend on its freshness, and goes down the drain entirely if even a single business entity definition changes. Spreadsheet owners are usually unaware of the change.
Totally concur that excel really ought to be for presenting data, however, people will always 'go rogue' and use excel for all bar the last purpose. This is because of a variety of reasons the main three being: wanting control, not wanting to bother IT, wanting it quicker than IT will deliver it. I would love to be in the ideal situation where every spreadsheet pulls from centralised sources and all the spreadsheets report the same figures with users being responsible and diligent in their knowledge of the data and acceptable manipulations but that is a utopia which cannot in practice be attained because we will never be able to gather 100% of all data required by the company whilst having users capable and interested enough to be trained to an intermediate to advanced standard.
A pragmatic solution that I've heard of people implementing is a certification for spreadsheets. If you don't pass a spreadsheet to IT for the formulae and everything to be validated and 'signed off' then people within the company will not accept you spreadsheet as accurate. This requires significant strength of will on the part of your executives to see this through, however, I think it certainly has a lot of merit. Plus, if you combine this with powerpivot and sharepoint, you can implement proper change control and access measures as well as refreshing spreadsheets appropriately.
On the point of data change management and it's impact on spreadsheet users - I think this applies to pretty eveything IT does. I've found that assessing the change on KPIs and data for any project or change is often not given enough time, and users are rarely 'reached' to the extent that post-change they feel comfortable with their MI. I think it is a very hard concept to communicate, and to ensure all relevant people are told about it and get it.
As the only MI person in a company of 150 employees and maybe 100 users, practically, vetting everyone's spreadsheet pre- and post- change for the flotilla of projects and changes occuring every day is unfeasible. The simple answer is of course 'ditch the sheets' but this a slow process without a significant portion of executives sharing the vision and allowing sufficient time away from their BAU requests and projects to go and enact the plan.
I'd be very interested in understanding where your organisation started BI-wise and a summary of how it came to be in the sensible situation it now finds itself in. Who were the key driving people? What were the impediments/obstacles encountered? What really helped? How long did it take?
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply