Stop Counting and Start Slicing!

You say you don't have enough data to start.  I say you aren't slicing your data enough.

Fixing things is hard.  It takes time and you never know what you should fix first.  Everything seems important.  You have to make distinctions between the various things that are broken in order to set priorities, and the most defensible and repeatable way to do that is to measure them.

Here are some thoughts on measurement.  I highly recommend Doug Hubbard's brilliant "How to Measure Anything", which inspired a lot of my thinking on this topic.  Gathering data is time-consuming and expensive, but you probably don't need as much of it as you think. 

For example, if you take only five samples, there will be a 93.5% chance that the median number will fall within the range of those five values.  No, you won't see the outliers or black swans, but you'll know quite a bit about your problem very quickly.  You can spot trends even more quickly with the Urn of Mystery rule.  Imagine you have a jar filled with two colors of balls.  If you randomly pull a single ball, there is a 75% chance that the majority of the balls are that color.  So quit saying you don't have enough data to know where to start.

But wait, you say, my shop is complex - we have lots of different-colored balls, not just two. Of course you do, and the solution is to slice the data thinner.  You aren't trying to fix everything at once, so why would you measure it all at once? 

Infopedian
[CC BY-SA 3.0 (https://creativecommons.org/licenses/by-sa/3.0)]

Pull out all your data and sort it.  Sort it by issue, by number of occurrences, by location, name, severity, model number, IP address - any attribute that can help you group it into smaller sets.  Grouping and ordering it differently can help you spot trends or opportunities.  When you finally get the right slice, you'll see a clear path forward.  You'll find a commonality that highlights a particular issue or set of related issues.  These are your quick wins.  Quick wins impress everyone, because it makes the team realize it's possible to make progress, even if it's tiny and incremental*.

At some point during this analysis, you'll probably also find that the 80/20 rule applies to your problem - ie, 80% of your problem can be solved by fixing 20% of your stuff.  Let me say that again: You can probably resolve 80% of your issues and only touch 20% of your stuff.  This is your big win, and big wins fuel morale.  Morale fuels momentum, and if you're not careful, you'll find yourself in the middle of a virtuous cycle.

So stop counting and start slicing!

* This type of data slicing is a simple form of OLAP pivoting