July 22, 2010 at 3:09 pm
Steve Jones - Editor (7/22/2010)
Greg Edwards-268690 (7/22/2010)
They found out that this 'creative' person had started with the SO Cycle Time detail, removing some data points, and presenting it as WO Cycle time. So for almost 2 years, they had been presentling progress towards a KPI using the wrong data. They found out that our rules and the data for WO Cycle Time met the requirements and the data was cleaning measuring what they wanted to see.
As a counter point, I've seen IT do this. The analyst says something, the developer mixes them up, and this gets built into an application wrong.
That is where a good process has both developer and user testing and signing off before something goes to production. If something gets missed, you need to review the testing procedure.
The person who inherited this used to be my manager. I had been one of the creative people in the past.
It never hurts to have 2 or 3 people validate, especially when you get a mix of operations and IT involved.
I've done my share of Ad Hoc when I was in Operations, and had my share become the prototype to production.
So I've seen both sides, and somewhere in the middle is usually a good spot. But I lean much more towards IT to publish the official corporate views to be used by all. But they have to be developed from user requirements.
Greg E
July 22, 2010 at 3:10 pm
l543123 (7/22/2010)
oh ya Power-Pivot, when our users are still struggling to use excel 😛
So, is anyone training them, or helping them learn how to use the tools effectively?
If not, why not be proactive, and push for this to happen.
Are you really good at using the tools? If the company won't pay, and you are good, why not set up a "Lunch and Learn" and let people come in and brown bag it.
It is a great opportunity to push for a positive outcome that will benefit the users and your organization.
Oh, and having taught a number of courses up to, and including the college level, you will be surprised at how much new stuff you will learn yourself!
July 22, 2010 at 4:30 pm
Here’s a question not many people consider about self-service BI - can it lead to bad decision making (a core reason for BI)?
Take this scenario - A set of experienced management accountants all with many years excel pivot tables and data handling all sat in the same office and reporting to the same FD (Finance Director) start their own self-service BI using excel and access. Yes, they got their initial answers quicker than a managed BI offering, however, a couple of months after conception the FD hauled the seniors in and asked about the huge discrepancies in figures he was seeing. Turns out that because it was self service the accountants were starting to use different techniques to produce figures that favoured their position. The FD ordered them to take a single approach, all was happy again, a few months later discrepancies started to happen again.
This actual case goes to demonstrate that in certain projects self service BI allows users to manipulate figures to their own benefit which can lead to misinformaion, discrepancies, conflict during meetings and generally impede or harm decision making. Whereas managed BI can ensures a ‘single version of the truth’.
Our solution was for me to work with a systems accountant for 3 weeks agiling a Trial Balance cube. Guess what, no more discrepancies and a happy FD and a happier set of accountants because they don’t have to contend with maintaining the necessary access databases.
There are many reasons you could hang the accountants failure on but technical experience is not one of them. In my opinion, in this situation self service BI was a big mistake. It gave users far too much flexibility to bend data to their own shape; it only takes 2 people to produce 2 different figures for revenue or margin or any other figure to cause confusion or lead to a bad business decision.
I do believe self service BI has a place in the business but the choice of project needs to be right and, again in my opinion, the business decisions made on the back of such systems should be considered with greater care than with managed BI.
July 23, 2010 at 7:10 am
I think that it is not an all or nothing type of deal. I saw the premise of the article was that IT can provide something quick enough to be viable (for a crazy IT schedule) and for the user to be able to see the possibilities and maybe get user requirements close to final before doing a full solution.
The politics involved in your story are a great example why. IT usually does not have a dog in the hunt, so there is no motivation to do anything but present an accurate picture as possible and to provide something that is actually useful.
July 23, 2010 at 3:17 pm
wildh (7/22/2010)
This actual case goes to demonstrate that in certain projects self service BI allows users to manipulate figures to their own benefit which can lead to misinformaion, discrepancies, conflict during meetings and generally impede or harm decision making. Whereas managed BI can ensures a ‘single version of the truth’.
Managed BI does not ensure 'a single verison of the truth,' and could result in IT inadvertently aiding in the misinformation, discrepancies, conflict, etc. we want to avoid.
My actual case. The VP of Finance of a company I used to work for came to me and asked me to extract some data for him manipulate in the process, and put it into an spreadsheet. The only reason I knew he was asking me to assist in a deceit was because I had been a business owner and had taken accounting course in college. Had I given him what he asked for there would have been a "single version", but it definitely would not have been the truth.
I offered him the raw data and told him he could manipulate it however he wanted, but that the other executives would also be sent the exact same data.
The more readily available the information is (which these tools help with), the harder it is for someone to misinform. It will also help in catching discrepancies, not make them more likely. As for conflict in meetings, I've seen more conflict when there was only 'a single version of the truth' then when it was available for more people to easily analyze. Also, not all conflict in meetings is bad. Frequently it leads to "teachable moments."
July 25, 2010 at 5:40 pm
I agree that it can be better for everybody to have access to the source data than to have the one source of data skewed to one person's point of view.
I also agree that it does not have to be an all or nothing solution. A managed BI solution can be provided as the single source to check against, then 'sandboxes' can be provided to any stakeholders who might have the business knowledge and technical savvy to use them. This provides some amount of checks and balances to ensure that the single source of data is actually correct. Some amount of challenge is healthy.
July 26, 2010 at 7:43 am
I think that self service and managed BI can and should live in the same house.
Self service is for the people with experience and understanding of the data and data analysis tools to create ad hoc and "hmmm what about this scenario" reports.
Managed BI is for regular reports that everyone agrees is sourced correctly and contains the facts that provide proper interpretation.
When the self service report becomes a regular thing, it should be audited and moved up to the managed BI platform. This will free time for the analyst to look into other scenarios rather than putting together and maintaining regular reports. My gut feeling is that this is where the relationship between self service and managed BI usually fails. People get comfortable doing their reports and are reluctant to give up control and move onto other issues.
July 26, 2010 at 2:45 pm
Steve, I agree with you entirely.
Besides I want to add something interesting: How many times had happen that we spend several days, creating something, just to know that it was not what the functional user wanted?
Make no mistake. I'm not talking about let them to do our job; but I sincerely think is a better approach to empower (a little bit) the final user, of the propossed solution; it will make them feel more comfortable with the project, and will make them feel an essential part of the project itself. Our task, in that case should be, help them to make the solution aligned with our best practices, tunning their queries, explainning to them the additional benefits, by running the right sentences.
But lets face it. Usually; the DBAs are not good making friends -We've been arguing with the developers since the very creation of our career-.
Personally, it's been a long journey, but after a couple of years, I truly believe that most of time, the user knows what he/she wants; what he doesn't know are the technical implications of it. But wait a minute: the technical implications are part of our job, don't they?
Computer Science is not like 30 or 40 years ago: users in their home have appliances more powerful than some of the servers that we have to administer today. Indeed, they don't know everything about IT, but in BI projects, we can convert them into allies if we give them some power with certain controls.
At the end of they, our paycheck is upon their satisfaction 🙂
July 26, 2010 at 8:45 pm
I like the point that we could start with something canned and develop that into something more specific to our needs. Let the users do a bit of it and then they may have a much better idea of what they truly want in the app. That could get us closer to what they want with the first release of the app designed for them.
Jason...AKA CirqueDeSQLeil
_______________________________________________
I have given a name to my pain...MCM SQL Server, MVP
SQL RNNR
Posting Performance Based Questions - Gail Shaw[/url]
Learn Extended Events
Viewing 9 posts - 31 through 38 (of 38 total)
You must be logged in to reply to this topic. Login to reply