Slashdot Mirror
How Do You Justify the Existence of IT?
bakamaki writes "I work for a small manufacturing company as a SysAdmin. My boss is a DBA. We are the only IT employees. He recently decided to record hours spent on his projects and then evaluate how much time the databases he writes save the employees. Then he translates that into a $ figure. He's asking me to do something similar but I'm kinda at a loss. It seems most of the stuff I do is preventative, IE care and feeding of servers and network infrastructure in addition to all the break fix stuff I do for the user base with their desktops. When in this position what do you folks usually do?"
There are 5 valid reasons for any business decision:
1. Legal: laws, rules and regulations
2. Contractual requirements
3. Positive impact to the bottom line by increasing revenue and/or decreasing expenses.
4. Quality of life issue for your customers
5. Quality of life issue for employees
You can look at things like backups and preventative maintenance as addressing both #1 and #3 as matters of risk reduction and business enablement. How much would it cost your company to not have its data? Or to not have access to it for 4, 8, 12, 24, or 48 hours?
Then you can look at the direct costing method: how many projects have you worked on, what were their budgets (capital and otherwise) and how much did your work contribute toward that?
All you do is stop doing your job and wait for everything to crash, then figure out how much money the company lost.
If he's the maintenance IT guy, he needs to take a different approach. He should show the costs with no maintenance IT guy, contracting out to another company, and then show his costs. If he's busy all day, it's guaranteed that he's saving them money. Contracting out will get the job done in the same amount of time, it'll just cost a lot more money. With nobody there, everything he deals with will still have to be dealt with, it's just that it'll have to be dealt with by people slower than he is and not as good at the job.
Overall, he should be able to show at least 40% savings over contractors and 70% savings over everyone dealing with it themselves. Almost everyone here's worked at a place with too little IT support and seen how it kills productivity, so this should be a fairly simple exercise. I suspect that his supervisor will have something to add to his presentation to cater it to the executives, but if the executives don't immediately see the merit of the report, then they would be hostile towards IT anyway and there's probably nothing that can be done.
It sounds like you do a lot of preventive maintenance. Now what you might want to look at here is how much income would be lost for the company if their employees sat around waiting for an outsourced tech to come and fix their systems, as opposed to having you on staff, PREVENTING those lost hours.
That's because you're taking technology for granted. If you weren't there, that technology couldn't be deployed to help people get their jobs done. Which means no servers, no desktops, no laptops, no networks, no printers, nothing, nada, zip, zilch, zero.
Not true. Many companies I've done work for hired me to come in and set it all up. The people that make this type of decision usually view a computer like a VCR -- they believe that, once setup, it should "just work" and not need any maintenance/etc. Plus, there's always "that guy" in the office who setup/built/installed "a computer for his buddy/mom/dad", and is obviously an expert. This guy will ultimately get tasked to do much of the work (I've seen this more times than I have fingers/toes).
Now all you have to do is compute how much it would cost to get common tasks done. Take handouts for a meeting as an example. Right now I'm sure that the employees type up the documents then print a few copies off the printer. Since we're talking about modern wordpressor technology, it would take them 2-3 complete, hand-written (or perhaps typewriter typed) drafts to develop the same document. Then they'd need to run the final document through the copy machine for the number of copies they need.
Again, this isn't what would happen. Every desk would have it's own printer (occasionally, there might be a network available printer a guy like me might have setup). They'll just print however many copies of whatever it is they need to distribute to people. If they also have an office copier, they'll print one copy and then make as many as they need at the copier. (You can use this paragraph as an answer to your "spreadsheet" excuse, also)
Now on to email. Remember inter-office memos? Back when entire mail departments were needed just to distribute memos between employees? Find out how many employees usually staffed these mail rooms. Add to this the cost of inboxes on desks, mail carrying equipment, space needed by the average mail room, and/or (if your company is really big) the infrastructure cost of pnuematic tubes.
You must live in a weird area if you've seen people do this. What I've seen are companies having their employees either sign up for a "company address" via hotmail/yahoo/google/whatever, or using their own personal accounts. They also install some common IM (MSN, Yahoo, AIM being the most common) to communicate with each other, if they have separate offices.
As far as "presentations" go -- Any companies that actually did this usually have a portable LCD projector, screen, and some sort of laptop (but this has been rare -- most of the companies I've worked for just don't hold these types of meetings). This type of hardware is fairly cheap, can be "locked in a closet", and doesn't require professional (or permanent) installation. To create the presentation? They'll just use one of the office computers mentioned earlier to create it. No Kinko's, no laying everything out by hand, etc.
I haven't even gotten into subjects like billing, reporting, and other data processing. Feel free to work out the cost of mainframes or (even worse) a small army of accountants and typists.
Billing/reporting? Peachtree/Quickbooks. Run, of course, from the previously mentioned office computers.
Out of curiosity, were you intentionally trying to make this out to be as difficult/primitive as possible? Would you happen to work for a company that makes infomercials? Reading over your post, I was reminded of that stupid infomercial about stacking tupperware where the lady opens her cabinets, starts flailing her hands about inside the cupboard, causing all the existing plastic-ware to fall onto her head. Sounded *exactly* like how you tried to describe how business would work without an IT person.
bork bork bork!
I budget for IT resources. Ten years ago, we were paying for 12 full time people at our data centers, but I never saw more than 6 in any month. A few of the guys were definitely worth 2x whatever they were paid. To find out what the other 4 were doing all day, we had them tell us the top 5 things that they did, then had them keep track of how many times they each did it in a week. They complained it would take too much time at first, so I had their boss-boss-boss tell them to do it.
Password resets was by far the largest waste of their time. Users had so many different accounts (20+) and used some of them just once every 4 months for very specialized tasks. Almost 10% of the users called for a password reset of some kind every week. We were running NIS and AD but that didn't help with application logins.
Centralized LDAP and automated password reset tool projects paid for themselves in just a few months. The hard part was in migrating all the web site authentication over and getting support in commercial products. We're still fighting that today, but users are down to 3 passwords at most, which is manageable - in comparison.
Printer Configuration fixes/changes. The 2000 printers were configured individually on each server, manually. Whenever a change was needed, these 4 geniuses would do it on the server that the end user complained about and none of the others. "I can't print to PRT010101 from SRV543." A centralized printer management project with trivial scripting and nightly replication to all UNIX servers ensured that any issue with 1 printer (move/add/change) is mirrored to all the other servers overnight. It took a few weeks to get the 4 "smart guys" to stop editing configs on any server other than the main print config server, however.
No disaster recovery testing except once a year on test DR systems with a 72 hour time limit to get it working or stop trying. We implemented most new projects with a fail over system in an alternate data center. Every week, we'd update DNS and fail over the production systems to the alternate data center. Yes, we could have used global load balancers for a similar end, but CORBA connections didn't like that back then. There was no need for DR testing anymore and we **knew** the DR system was current since we tested it every other week. Everyone sleeps better at night with this one. As systems are upgraded with new hardware, we implement DR/Test systems for them too. There's a big difference between "hoping" and "knowing" DR works. It is also an easy back out plan for any system or application updates.
Basically, figure out where/what you spend most of your time doing then optimize that effort/process. Work to the 2nd and 3rd and 4th most time consuming items and optimize each of those in turn. If you spend an hour a day doing something and can reduce that to 10 minutes, you've saved 4 hours a week so you can do more interesting things. Assume you cost your company $100/hr - that's $400 per week in lost opportunity costs, $1600/month, almost $20k/yr - just this small efficiency improvement can pay back BIG.
There were other stupid things that I can't recall immediately, but it all started with counting the number of password resets. Perhaps you've already solved those items. Perhaps you are still walking to each desktop to work on them? If you could remote into the desktop to fix things, you'd save a bunch of time and limit the "hey, I've got a problem, got a sec?" issue. Or perhaps you have users who install software that break other software. Locking down business computers to prevent stupid user tricks can be a huge time saver.
I've been doing on-call support for years now and I've found the best way is a subscription service - basically a monthly amount so that I'm available when they call. I service primarily dental offices and base the amount on number of workstations and servers, what applications they are running, and also what type of response time they are looking for. Someone is physically onsite at least once a month regardless - this is what some other small IT shops miss with these types of arrangements. Clients don't like getting a bill and never seeing a warm body onsite.. no matter how good your remote support setup is.
Emergency onsite calls are charged extra at a fixed rate - most other stuff can be handled during the monthly visit and/or remotely. Very few clients want an 'all-inclusive' arrangement where they pay a fixed amount for 'unlimited' service.
We used to do break-fix but found that it was much harder to retain clients long-term. Billing is a pain and sometimes difficult to justify to the client. We also found that the key to staying afloat was to 'cull' your client list every year - drop the 10% that never pay on time, are a pain-in-the-ass, and so on. This frees up time to find clients that you do want to keep.
Keep in mind that my experience is limited to SMALL businesses - biggest client has 45 stations and 3 servers.
Please stop APK.. you're only hurting yourself.
I dunno... ask your neighbor? Seriously, there's gotta be a number for your helpdesk!