Cloud, Technical Debt, and the Quantified Self

April 2, 2013 Nick Hamm

I’ll admit it – I didn’t do a very good job keeping up with my eating habits and personal fitness routine last year. I could blame it on work, travel, small children or probably several other “justifiable” excuses, but the only thing I know for sure is the number that pops up on the scale every morning when I step on it – and it has been increasing at a steady rate.

I recently read this fascinating discussion on technical debt with thought-leaders Ward Cunningham and Capers Jones. This is a must-read for anyone who develops software or manages a technology program. It’s full of some very interesting stats and metaphors that will make you think about how you manage your projects. Yeah, it’s long, but it’s worth your time.

Today many enterprises are working to pay off their legacy technical debt by moving to cloud platforms like Salesforce, Google, Workday, and AWS. It’s a great feeling to end of life a hairball entanglement of systems and consolidate on a common platform with no hardware or software. One of the great benefits of cloud is that you already start much lower on the technical debt scale because you have no hardware or software to manage.

But buyer beware. Even though you may be starting lower on the overall scale, cloud is not immune to the build-up of technical debt. In fact because of the ease of customization and the ability for your team to focus on value-add requirements instead of platform-related minutia, technical debt can actually grow at an accelerated pace in the cloud as opposed to older, less flexible solutions. And without proper measures in place, before you know it you’ve marginalized the benefits of moving to a cloud platform by stuffing it full of every backlog item from the past 3 years.

Unchecked, cloud can be like putting a starving business in front of an all you can eat buffet. Without the right discipline it takes very little effort to consume thousands of calories very quickly. And before you know it you’re 10 pounds overweight and can’t run as quickly or jump as high as you thought you could.


In life and in business, what gets measured gets managed. The quantified self movement has changed the way we behave by giving us a new way to hold ourselves accountable. It’s also allowed us to have deeper insight into our actions and performance. At the core of quantified self is self-tracking and self-knowledge, which gives us the data to make better decisions about our diet, exercise, sleep, and other factors that affect our health and performance. These same principles apply to enterprises implementing cloud solutions.

The two key ingredients in any successful technical measurement program are content and context. Having the data is the first key element. How many lines of code do we have in our environment? At what rate are we adding code to the environment? How are we doing in our balance between configuration and customization? These are just a few of the questions that you need to be able to answer. But it’s only half of story. Which business processes have you implemented? How rapidly is your industry changing? What percentage of your staff is dedicated to building and maintaining your environment? Contextually understanding the factors specific to your implementation as well as how your data points compare to benchmarks from companies similar to yours gives you the intelligence to make trade-offs, push ahead, bolster your team, or step back and refactor.

Appirio has been collecting Cloud Metrics with the most advanced enterprises for last several years. The insights on the health and performance of these orgs have been tremendous and have helped us better serve the interests of our clients. Historically this has been part of how we manage programs at Appirio, but today we are encouraging enterprises to take the initiative to monitor and manage these metrics as a part of their own cloud program management whether or not you already work with Appirio in any other capacity.

For example, the range for lines of code in one cloud application for our top 10 benchmark accounts is 92,000 – 231,000, while the median sits at about 15,000. In the case of this single metric, bigger is not necessarily worse – it’s the context that explains the significance of the data. What the data can tell you though is that you need the resources to support that amount of code. It tells you that you have a certain level of complexity at which you are operating that may have unforeseen side effects. And as that number grows or shrinks, so should your operating assumptions. Without this data you’re only guessing, which can be like flying a plane through heavy fog without instruments.

 

I’ve started measuring and paying down my health debt this year, and I’ve noticed a positive difference already. It feels good to look back and see the improvements that can be made in even a short timeframe. I’d encourage you to do start doing the same for your enterprise’s technical debt before it’s too late.

Nick Hamm is member of Appirio’s Technology team and is responsible for Appirio’s Cloud Asset Library. He is a Salesforce MVP and has helped over 200 companies across a wide variety of industries transform the way they do business by implementing cloud solutions.

Previous Article
7 Ways to Use CRM to Enable Your Sales Team
7 Ways to Use CRM to Enable Your Sales Team

Sales organizations can succeed solely with good sales reps and solid leadership driving them in a focused ...

Next Article
Reimagine Talent Planning
Reimagine Talent Planning

In a recent post we talked about the importance of creating an agile talent management strategy in a new gl...