Yes you can measure engineering | Jade Rubick - Engineering Leadership<br>Available for advising, and individual or group coaching.

I’d like to share my favorite way of measuring the value delivery of engineering organizations. It’s something I’ve used at a lot of companies. I call it the ValueSum metric .

Why the ValueSum metric?

I like this metric because it has a number of advantages over anything else I’ve seen:

It is a cross-functional measurement , meaning it measures engineering, product, and design. The measurement incorporates how each of the functions are working separately, but also shows how well you’re working together.

It’s less easily gamed than most other metrics. And many of the ways you might game it aren’t harmful.

It proxies the value of the work delivered more than anything else I’ve seen.

It can be applied retroactively , so you can compare changes you make in the future with things that happened before you even started making the measurement.

It has some nice hooks for retrospectives and learning , so you can improve as a team.

Yes, you can measure if AI is increasing the value delivery of engineering ! You’re measuring the whole value delivery.

What is the ValueSum metric?

The basic idea of the ValueSum metric is that you add up everything that you deliver, and each item is multiplied by how valuable the work was.

The bigger the score, the better.

Let’s say you have three items that you deliver in a one-month period. You score each item using a rubric and you end up with a table that looks like this:

DeliverableScoreNew pricing tool6Automatic invoice generation8Foreign currency support3ValueSum for time period17<br>Why is that score useful?

There are a couple of things to know about these scores:

The faster you deliver things, the more things you deliver. And then the higher score you have.

The more valuable the work you deliver, the higher score you have.

So essentially this is a metric that captures value delivered over time .

At its heart, a lot of product management approaches prioritize based on value over time. This just extends that same concept to how you measure afterwards.

How to implement ValueSum

It’s straightforward to implement the ValueSum metric:

Create a scoring system and a tracking spreadsheet.

In that spreadsheet, create a list of all the things that have been delivered recently. I usually will go back to the beginning of the year or the previous year.

Have a product leader score each item, based on a rubric. This gets you a baseline.

Have an engineering leader or tool add how long each item took to deliver. (This isn’t strictly necessary, but can help identify items that were most valuable per unit time).

Put in place some sort of process to ensure you add new items to the list and score new items periodically or continually.

Talk about the metric in meetings. Put it on your dashboards.

That gives you the basics. Once you have the basics in place, you have a lot of sneaky leverage.

Sneaky leverage 1: can compute ahead of time

Is ValueSum a leading or lagging metric? Technically it’s a lagging metric, but you can also use it as a leading metric.

If you have a good work culture, an organization that understands ValueSum will want to increase that number over time. We want to get better at the delivery of value, right?

A thing to introduce is the idea of scoring the value of the items before they are delivered. (Product might already be doing this).

And you can loosely estimate the effort ahead of time.

But most importantly, you can have discussions about how to improve the score: can we make it more valuable? Can we make it take less time, or be delivered more incrementally?

Involve the team in these discussions, make them a natural part of the team’s vocabulary.

Profit.

Sneaky leverage 2: can retrospect and learn

Your scoring system is going to be imperfect. People never estimate the value very well. Or even score very well. So…

Three months or six months after each feature is delivered, have people rescore the features based on the value you’re now seeing.

Your perspective will be different, and it will probably be sobering.

But it will also help people score better in the future.

This can help the organization develop the discipline to ship software people actually value .

You might keep both values around, so you can see ValueSum and RetroValueSum together. That will give you an idea of the spread between what you predicted and what you evaluated retroactively. Over time you probably want to see that spread go down (but don’t optimize for it).

Sneaky leverage 3: can validate early

Product development organizations are often pretty bad at validating (or invalidating) ideas.

They can be either too slow, or too ineffective at validation.

ValueSum incentivizes validating ideas early. Why? You don’t want to invest a lot in something that isn’t valuable.

So, create small projects to prove or disprove something will be...