Skip

Measuring IT success
March 9, 2011 4:52 AM   Subscribe

How do I measure the success of my IT department?

I work in the finance dept of a medium sized company. At any given time, our IT dept is working on about 40 projects for the rest of the company. They've asked me to help them evaluate their success in quantifiable ways. I have no experience in this type of analysis but I'm happy to learn. Can you share any experience you might have in this area and point me to some resources to guide me?

I basically need to come up with a very condensed version of this for my company.
posted by yawper to Work & Money (8 answers total) 6 users marked this as a favorite
 
Like any other division of your company, the metrics you need to use are going to depend on your type of business. If your business processes orders over the web, then the reliability of your website may be very high on your list of performance metrics. If you process orders over the phone then perhaps a few days of website downtime isn't as critical.

You could start by identifying the natural groups in your company, take a small sample of the individuals (not just management) from each, and interview them about how they use technology in your company currently, how it worked in the past, and where they see needs for improvement.

It's worth doing the same on the customer side, if possible. This would give you a better idea of where your company relies on specific technology, which would enable you to design some performance metrics around those areas.

I'm afraid there isn't going to be a generic answer.
posted by odinsdream at 5:40 AM on March 9, 2011 [1 favorite]


Agree on the no generic answer. You want to measure Service Impacting Events, you want to know how many of those were preventable (and how they were preventable), how many SIEs were prevented or mitigated by monitoring or other proactive actions. Of course, you want uptime/service availability. Do you do helpdesk? You want tickets, mean time to resolution of tickets, ticket source (user error/network/virus...). Cost, you definitely want costs. You need to set a price, even if it isn't actually billed, on the cost of time spent for a helpdesk issue, or time spent setting up new users. This is really a huge area, and something that you will refine over time. And here I am, pontificating before having coffee... good luck, feel free to MeMail me if you want more coherent thoughts when I am awake!
posted by kellyblah at 6:02 AM on March 9, 2011


Do a search on SQM - Service Quality Management.

It's going to take some time from the beginning because before you can quantify their success, you need to know what to quantify. This is not a process you can rush if you want quality measurements.

It looks like this page gives a good overview and lots of links for you: http://dougmcclure.net/blog/service-quality-management-sqm/
posted by jillithd at 6:24 AM on March 9, 2011 [1 favorite]


Seconding taking time to get the metrics right. I was part of a group that was supposed to determine success metrics / value metrics for IT projects. We were given a short deadline, and the initial round of metrics were crap. Then we took 4 months to clean them up and came up with metrics that were understandable, measured relevant improvements, and could be used for years.

If you can get numbers on customer-facing areas like keeping your website up and accepting orders, that's great. Another area you could explore is efficiencies from internal projects - for instance, time saved by eliminating or simplifying steps in a process. If the process used to be "print form > fill out form > fax / mail form > receive form > record information / file form" and now is "Access form online > fill out and submit form > receive form with information already saved electronically" someone at your organization can probably observe, estimate, and calculate the time savings per month, which can be turned into a dollar figure based on the amount of employee time spent on the old process.
posted by Tehhund at 7:04 AM on March 9, 2011


Another thing to think about - I'm a member of ASEM (American Society of Engineering Managers) - and I know that the masters program at my local university had internships with local businesses specifically to do this kind of thing. It looks like you are in Canada, so maybe contact someone at CSEM to see if there are similar programs in your area.

If you can get cheap-ish labor from very smart, driven, and interested engineering management students who are training to do this specific thing, how can you really lose? Plus they get the experience from putting what they learn into practice. :)
posted by jillithd at 7:33 AM on March 9, 2011


Hi Global IT Strategy and Governance Manager here.

From above, it's a murky world you are about to enter. Measuring the success of IT depends what you want to measure ---- and more importantly ---- who is doing the measuring.

An easy way to start is to split baseline operational cost, i.e. what is used to keep the lights on so to speak - helpdesk, maintenance, staffing etc. and the projects. For example if Bob the IT director has a 1M yearly budget including projects. After breaking it down, you will find projects comprise 30% of that 1M. 70% is the lights on.

Then analyse what services are being offered within the operational budget. Again, helpdesk, licensing, maintenance, hardware costs, etc. Do the same for the project portfolio.

Now you see where the money is going, it's time to make some metrics. And herein lies the secret to getting KPIs right in the business. Don't create metrics alone. Your best bet is to propose a working group that comprises of business owners and senior IT managers. For example, one of the most successful groups I have ever worked with took an executive level overview (HR director, CIO, CFO, COO) to determine a top down look at what the company wants to measure and then asked the senior managers of each division to actually develop and approve the metrics. This way the organisation takes group responsibility for making sure what you measure is valuable to the organisation. Never fails and it is fun to watch from the sidelines the politics that inevitably become involved.

You will need a project coordinator to do this correctly and add value to the business. If your IT department is looking to get feedback - do it the right way the first time. The group will be amazed at what comes out and the business will feel responsible for the use of the money/performance. Set deadlines for review and approval. Plan to take 3-6 months depending on the size of your company to get the metrics up and running. Make sure business and IT owners receive the information.

Set a schedule for metric review using the above executive group and lower level working group. You will be amazed how quickly people will start to examine the spending and performance placed on IT with a focus to improving.

In the real world, our executive team meets quarterly to look at the big picture and make sure things are on track. The senior management team sits once a month and has the power to implement controls/changes etc.

If you need more information, memail me and I can find you some starting points. Did I mention do not do this alone or just with the IT department? Get your metrics in as a project the business has to take responsiblitiy for.

Good luck and again, if you need help or some starting points, glad to help.
posted by Funmonkey1 at 7:33 AM on March 9, 2011 [2 favorites]


There are various ways to measure the pure IT/systems performance, it's those projects that can be a very ethereal entity. They should be estimated, proposed, budgeted, scheduled, tracked, signed off, and then Customer-satisfaction-surveyed just as if it was an external project for which you would bill. That way you'll have the data to measure (and in my experience this is a fantastic way to curb the "hey, just real quick" items that turn into multiple-week time-sucking monsters). If you're using or thinking of implementing a support ticketing system to handle the helpdesk side of things, be warned that most of them are crap for long projects because they keep trying to penalize or don't have the features for longer multi-step projects.
posted by Lyn Never at 7:38 AM on March 9, 2011


Wow...thank you for all these fantastic answers and for the offers to help over memail. You guys are great. It seems I have a big project ahead of me...
posted by yawper at 9:12 AM on March 9, 2011


« Older Looking for cheap accommodatio...   |  How do I get the cigarette sme... Newer »
This thread is closed to new comments.


Post