Some would argue that it all started with Lotus 123 back in the 80s. When the early spreadsheets emerged and users seized on them as a way of solving their own information requirements, IT's control of the organisation's data started to gradually slip away. We could put all of the security and integrity checks we liked into …
It's still a nightmare
People write a report and then fail to update it as the business changes (think long where clauses which include various rows and aren't changed as new types are introduced). The original writer leaves and someone inherits the report, and has no idea how wrong it might be. These things become part of the monthly reporting cycle i.e mission critical.
We are currently trying to identify all 'critical' reports which will then become 'owned' by IT. Any changes required will be managed by IT and the business folk will be able to run the reports but not change them.
Extra work for IT yes, but we can then make sure the SQL used is efficient and the results accurate. It will end calls for help on reports that no-one has ever seen before which can take a large amount of unscheduled time to resolve.
Business benefits from BI
Business Intelligence is a beautiful thing but so few businesses - my own included - understand how to make it so. There's 3 problems as i see it:
1. Soooooo much data
2. So many people thinking they need that data
3. Getting that data into something useful
To give detail on that, the amount of data a business generates is something of a self-perpetuating ritual. Everyone thinks they need to know everything about everything. But clearly they don't.
I'm currently working on a major BI case across four business units - all baying for data. The rule of thumb I apply, and indeed my mantra is:
' How will you action that peice of data into something that delivers a business benefit.'
That throws a lot of people. But if you start off with the major list and then work it through into:
Data X will enable us to deliver a 0.1% improvement on ROI via department Z and - person A will be responsible for that delivery.
That sorts the wheat from the chaff. It keeps everyone focussed and makes them realise that, half the time, they just want data to feel important.
Knowledge is power - but only if you know what and how you'll use it.
It seems that every "power user" I know has a different set of tricks, and that if you put more than three of them together in a room, somebody is going to say "I didn't know that". My wife constantly amazes me with her skills in Excel, but I surprise her with my Powerpoint. And so it goes.
Most "power users" are too intent on showing off their knowledge of the sublime to actually get the message across succinctly. In some fora, having that extra pizzazz helps, but often the recipient is either stunned by the artful presentation of indecipherable data or confused by the obfuscating techniques of data that is clear. Even a powerful, artistic Powerpoint presentation, done with esoteric tricks, can confuse the audience with a presentation where the media overwhelms the message.
The "power users" are frequently too embarassed to admit that they either don't know what they are doing or have FUBARed something, with no tenable backup. It's the guy who sees some neat trick and wants to do it, but won't ask the "power user" for fear of looking dumb. So guess where he goes.
As a developer in a large insurance company, everything I write is according to well defined specifications and there is a pretty strict testing and release process (some might say overkill) before anything goes live. I know damn well though that there are a lot of Excel spreadsheets used to run the business that have complicated formulas and home grown macros in them that have never been even reviewed by someone other than the person who wrote them, let alone properly tested. It's no wonder mistakes get through and they sometimes end up with numbers that don't make sense. Worse still, when the answers are just a little bit out - but not enough to be obvious!
The other side of the story
Interesting article. I suppose I would qualify as a "power user" myself, and I admit that I do tend to sort myself out independently of our IT people, but that's because requests for help are generally greeted with a cost quote and an absurd timescale for delivering anything. I shouldn't have to spend the time I do getting data out of different systems and combining it to support basic needs, but I don't have much choice. Its not that IT is unhelpful, some of the support guys are really good, but they are only allowed to work on things that are in their plan.
It's all about the (l)users...
I've been involved with the implementation of over 100 very large data warehouse (20TB+) around the world. And, yes, the problem of "power lusers" does occasionally rear it's ugly head. However, in most organizations I've been in, the problem has really been well contained through three very simple (albeit politically charged) strategies.
(1) Use monitoring tools to limit "unauthorized" use of resources. This is fairly straight-forward. Most data bases have tools that allow restriction of resource by user or group, both static (restricting the amount of work space allowed for intermediate answers in a query) and dynamic (monitoring the progress of a query in real-time and cutting it off when it exceeds a set of thresholds). This allows even "casual" users access to various data sources via ODBC or other "free-form" tools, but limits their impact (and time wasted) to manageable limits.
(2) Provide a "liaison" resource for each group that has a business requirement to access warehouse data. It is amazing how much this simple addition can do to both limit users abuse of resources AND provide real business benefit for the group. This "liaison" person is the "power user" for the group; however, they are cross-allocated to the DW support team and have responsibility for being both the advocate and enforcement for their group. Use of the "liaison" person gets the majority of casual users back at their desks rather than playing with tools and code; and it provides a focal point for getting ideas discussed with other liaison members and DW support.
(3) Provide the users with tools and data that match their business needs. This is the major "sticker" in most organizations...and the reason that BPR is such a lucrative profession to be in. Many of the problems outlined in the article are, as noted, due to data no longer being relevant to their business functions. It is rare that organizations truly explore the impact that a business process change makes in their data requirements.
However, in this last scenario, having a "liaison" person on the team also helps catch and prevent this last problem from occurring. Or at least identifies the problem so it can be properly ranked for impact and cost.
The bottom line is, of course, cost. If an organization is having problems with "power (l)users", it generally indicates a larger problem in corporate management and governance. This can be either a boon or a curse: as a consultant I use this as one of the metrics to decide if I will take a contract or not.
Other people's broken systems
I am currently wading through an extensive set of MS Access databases designed to pull custom reports from an exported subset of an SQL database. It was written over five years ago. Since the office is part of the military, our personnel turnover interval runs about three years or less - sometimes much less. As you might guess, the original programmer failed to provide ANY documentation to his databases and code. The second culprit, who rewrote portions of it in response to changes in the SQL source files, added a few comments here and there regarding his changes, but the comments are inconsistant, and his changes are obvious patchwork - and of course, his comments only cover portions of the VBA code, not the actual tables, queries, and forms. Now I find myself in that "Power User" position, trying to deal with a badly broken system that is so integrated into our daily processes that it cannot be scrapped. It has taken me three months of dedicated effort to develop a moderately useful set of documentation, and to begin to comprehend the overall system well enough to make changes and correct errors. Worse, other bright boys have added their own modifications - making a daily copy of our intermediate database in a separate location as their own source for a different set of reports. This means that every change I think up must be considered against the results of three completely separate output streams.
I don't know how much IT staff time I'm wasting - not very much, since I know already that they don't support non-standard applications - but I know how much of MY time is getting sucked into this black hole. On the upside, though, I'll only be here a few more months. And at least MY successor will have a thoroughly hyperlinked Word document that details every table, query, form, and module - he won't have to start from scratch.
I used to work for a major UK financial company in the internal IT consultancy, we also worked as 3rd/4th line support. It was a relatively common occurance for a support call to end up with us that went along the lines of:
"...Customer's department has an excell/word macro, that is essential for their business unit, has been developed, by a temp who has since left, and is password protected. The macro now needs to be updated to clear x/y/z bug, can you sort it out for us..."
To which the answer was usually it'll have to be re-written by our programmers because they didn't comment their code/the code is all over the place/we can't crack it's encryption. This usually results in the business having a fit at the cost and time required for redevelopment.
We also had massive problems with our Acturies. This is not uncommon. One particular incident springs to mind: The acturies went out and, without consulting the IT department, purchased a £12k piece of software, which, unbeknown to us, they ran on their own farm of workstations. The business continuity people noticed this and instructed them to get the system into the datacentre, at this point they approched the IT department. The problem was that we couldn't allow workstation class hardware into the datacentre, but the software they had got was so badly written that it actually worked less efficiently on a Xeon processor. In the end we had to compromise and put shelfes into racks so we could load them with workstations, which then started to fail because of the heat produced by their power supplies....
Users are gaining more power with less experience
The ability for a user to access the data they need is becoming easier everyday, with programming falling by the wayside of automated feeds and "web 2.0" content; this proves problematic in the respect that users have less idea of what they are actually doing, and are more likely to end up in a mess, which they expect us to sort. The way IT has become has allowed users to become more involved, which is a positive, but is nieve of their inexperience, which is a negative. Copyright will soon hit hard along with a wave of lawsuits, and I intend to stand on the side of the professional, not the user.
That's a bit bloody idealistic.
That sounds crazy.
I'm sure that the average user is badly served by the IT systems. I've worked on IT reporting systems for a decade, and many of them are flawed by design (or a bad technology has been chosen).
Having said that - the average user also seems to be stuck in somewhat of a rut. Given the best reporting tool that money can buy, many would simply sit at their desk exporting data to excel repeatedly so that they can print it, stick it on the wall, and mail it to their senior management on their weekly and monthly due dates.
So what I'm saying is that IT is stupid, and end-users are just as stupid. IT produces duds mostly, and given a chance, end users will generally simply reproduce the same numbers with less effort - rather than perform new analysis.
I work for one of the top five largest banks in the world, we recently integrated another bank we own into our systems. We stopped counting their piddly little access databases/visual BASIC/macro apps at at 5000, there was no way that we were going to be able to deal with them, so we just stuck them all onto a massive SQL server. The plan is to deal with them one at a time when they fail. Erk...
Piddly becoming business critical
I think most people here would be able to recollect a time when they have acted as a "power user", or when they've had to unpick someone's previous work.
I think the important thing to remember is that nobody ever aims for their excel/access/vba masterpiece to become a business critical tool. They start out as a tactical means to an end, and if anyone ever suggested getting full funding/infrastructure/licences for the strategic solution (data warehouse, sql server, whatever) the answer is always a quote in the scale of thousands, compared with "free". The problem arises when people become dependent on the tactical solution and it becomes business critical. The fact that it's a gradual change means there's usually no tipping point that anyone can identify, or it's too late when we do notice.
I think it'll always happen to some degree, the old "sledgehammer to crack a nut" comment will be brought up instead. It puts it in perspective to think that for every business critical excel sheet, there are 100's that never got that far, which we'd never consider putting on a data warehouse!
Enthusiastic power users still a danger?
A friend of mine put it very well:
it used to be
"Mom, Timmy is in the backyard with scissors!"
"Mom, Timmy has Daddy's chainsaw in the living room!"
Depends on the organization
The only place I have worked at which power users were both empowered and well directed was an optoelectronics manufacturing firm with strong and forward-thinking management. However, even at this firm, it was entirely likely that any task I performed (as IT-Production liason) would involve an endless maelstrom of redirection from one manager to the next because no one knew the whole organization. In order to complete tasks, I deviated from instructions, which breaks ISO9001 compliance. So, here I was part of the IT support team, but effectively a troublesome power-user with respect to the rest of the organization.
Every other organization I have worked with suffered (or continues to suffer) from a lack of over-arching managerial understanding. Basically, as an organization, we have no clue what we're doing. From my point of view, this stems from a lack of focus on the goals of the organization (to make money?), an inability to distinguish (and define) behaviors that are beneficial and behaviors that are detrimental to the organizational goal (including the security aspects of utility verses risk), and a lack of willingness or inability to define appropriate levels of service according to the organizational goal and according to the resources available.
The result is always randomness, duplication of effort, inefficient investment in resources and purchasing practices, and a failure to meet the goals of the organization - in a word, waste.
What power-users don't seem to understand is that not everyone is a power user. Given the same tools, perhaps one in ten, or one in fifty users, will use them very effectively. Adding more tools doesn't help, but may actually decrease the ability to provide effective support .
What IT doesn't seem to understand (in cases I have seen) is that in their/our role of empowering creative work, facilitating productivity, and protecting the organizational data, appropriate levels of access and appropriate levels of service must be provided. Saying "No." without organizational backing is ineffective. Refusal to evolve is ineffective. Endlessly requesting more money for IT budgets without providing an understandible cost-benefit analysis is ineffective.
It helps if the organization is directed by people who have a solid grasp of the difference between science and science-fiction.
Basically, we're doomed...
- Mounties always get their man: Heartbleed 'hacker', 19, CUFFED
- Batten down the hatches, Ubuntu 14.04 LTS due in TWO DAYS
- Samsung Galaxy S5 fingerprint scanner hacked in just 4 DAYS
- Feast your PUNY eyes on highest resolution phone display EVER
- AMD demos 'Berlin' Opteron, world's first heterogeneous system architecture server chip