Leaving Software on the Table

Leaving Software on the Table

Read on My Site

Developer inefficiency drives hundreds of billions of dollars of lost software output annually.

The sum is striking but also predictable in light of data showing we don’t even come close to maximizing developer productivity.

Mountains of technical debt and poor development practices burden and bog down developers. We lose billions as a result.

The Developer Productivity Manifesto has three parts, this is part 3:

In part 1, I talked about falling developer productivity and how spinning the developer productivity flywheel can counteract this trend.

In part 2, I argued that more developers won’t solve all our software engineering problems.

I’ve thrown around a lot of equations, literary references, and even some memes. In this third and final piece, I’ll talk dollars and cents, quantifying the impact of lost developer productivity and how much software we’re “leaving on the table” as a result.

(Technical) debt to (developer) GDP

In part 2, I made the case that software maintenance is not sufficiently value generating to cover engineering salary expenses. In other words, busy work doesn’t work:

According to one analysis, an engineer engaged in purely non-innovative activity destroys nearly $600K in employer market value. On the other hand, the average engineer, working on a combination of maintenance and innovation activities, adds $855K in market value to their employer.

In a global, cross-industry survey, Stripe asked developers why productivity was lower than it otherwise could be. Maintenance of legacy systems / technical debt took the top spot:

Source: The Developer Coefficient, Stripe

Even developers themselves see maintenance work as unproductive. Asking them to quantify this reveals that the typical developer spends 13.5 hours per week addressing technical debt:

Source: The Developer Coefficient, Stripe

Add to that another 3.8 hours per week fixing “bad code” (debugging, refactoring, etc.). That totals 17.3 hours per week spent fixing the past rather than building the future. The typical work week among the surveyed was 41.1 hours, implying that a full 42% of developer time is lost to drudgery:

Source: The Developer Coefficient, Stripe

I like to call this technical debt to developer GDP. While some level of technical debt is unavoidable, it eventually becomes an unbearable drag on software output and developer productivity. It eats up engineering time leaves little for generative development work.

Developers could be 46% more productive

In another question, Stripe asked developers to rate the productivity of their engineering teams on a scale of 0–100%. The average response? 68.4%:

Source: The Developer Coefficient, Stripe

Said differently, the average developer could potentially be (100% — 68.4%) / 68.4% = 46% more productive that they are today, nearly 50%.

With a 41.1 hour work week, such a productivity boost would be the equivalent of an additional ~19 hours of productive development work, enough to completely compensate for all that time spent on technical debt and bad, buggy code.

A $425B dollar bill on the ground

Based on the survey, Stripe conducted a back-of-the-envelope calculation multiplying estimates of the value generated by software developers around the world with the estimated productivity losses to arrive at an estimate for the global GDP lost due to software developer inefficiency.

I take issue with some of the assumptions, but the calculation is illustrative regardless. Assuming $900B of aggregate developer GDP and 31.6 percentage points of productivity lost suggests a $300B annual GDP shortfall:

Source: The Developer Coefficient, Stripe

This is an underestimate, in fact. Stripe’s math assumes 100% — 68.4% = 31.6% lost relative to existing productivity, but as I showed above, it’s in fact 31.6% / 68.4% = 46%. With this efficiency loss relative to maximum productivity, GDP lost to developer inefficiency grows to ~$425B.

So our failure to maximize productivity is losing us hundreds of billions in software production. Not exactly chump change.

SUM()-ing it all up

To end, I want to return to where I started qualitatively, my goal to increase total software output, and quantitatively, the decomposition of software output into developers and developer productivity, and finally connect the two perspectives.

How much more software output could we get with more developers and higher productivity?:

Let’s start with developers. I tend to be skeptical of perennial “shortages” unless there’s some specific, identifiable reason for it. Regardless, estimates suggest there’s a ~3M shortage of software developers globally, with 1.4M unfilled computer science jobs in the US alone.

To make the numbers easy, let’s round up Stripe’s estimate for the global software engineering labor force from 18M to 20M (I’ve seen other estimates in the 20–25M range, so this feels reasonable). That would imply a 3M / 20M = 15% developer shortage at current levels of demand.

So we have a 15% developer shortage and a 50% productivity “shortage”. Multiplied, that yields a massive 73% potential gain in software output:

Said another way, current software output is only about 100% / 173% = 58% of what it could be.

That’s $670B of software we’re leaving on the table.

This calculation is admittedly simplistic. One could argue that we’d need fewer developers if they were more productive. I push back on that — there’s so much we’ve yet to build, and more developers working more productively would unlock entirely new opportunities that we haven’t had the capacity to explore or can’t even yet imagine. This would further spur developer hiring and employment, another flywheel.

We can do better

In this series I’ve made the case that by investing in technical tools for technical people, we can reverse the trend of declining developer productivity, spin the productivity flywheel in the right direction, and see massive gains in software output as a result.

As I’ve hopefully made clear, we can do A LOT better of a job maximizing developer productivity. I hope you’ll join me in my mission to do exactly that.

Thanks so much for reading this — I hope it resonated with you. If it did, please share that feedback!

Follow me on Twitter and subscribe to my monthly essays here

Copyright © 2021 Who is Nnamdi?, All rights reserved.
You are receiving this email because you opted in via my website.

Our mailing address is:
Who is Nnamdi?
2200 Sand Hill Road
Menlo Park, CA 94025

Add us to your address book

Want to change how you receive these emails?
You can update your preferences or unsubscribe from this list.

Email Marketing Powered by Mailchimp

Older messages

The Developer Productivity Manifesto — The Flywheel

Wednesday, March 31, 2021

Developer productivity is falling. But it doesn't have to. The Developer Productivity Manifesto — The Flywheel Read on My Site I've been obsessed with developer productivity for a long time. As

Robinhood Traders are Last to the Party 🎉

Monday, March 1, 2021

Robinhood users get front-run by other retail traders Robinhood Traders are Last to the Party Read on My Site The recent Robinhood and Gamestop / Gamestock / Gamestonk fiasco shed a light on some of

There's Nothing Magical about the SaaS Magic Number

Friday, February 5, 2021

Sales and marketing spend drives much less revenue than you think There's Nothing Magical about the SaaS Magic Number Read on My Site I've never liked the SaaS "magic number." It

Product-Market Fit is Lindy

Friday, December 18, 2020

The longer you search for product-market fit, the less likely you will find it Product-Market Fit is Lindy Read on My Site The longer you search for product-market fit, the less likely you will find it

Why We Will Never Have Enough Software Developers

Wednesday, October 28, 2020

Developers are dropping out of the profession in large numbers Why We Will Never Have Enough Software Developers Read on My Site We will never have enough software developers. Developers are dropping

LinkedIn gets another chance, Lordstown gets the Hindenburg treatment, and Bob Redford's horse ranch (one of them) is up for grabs

Monday, June 14, 2021

The news you need to know June 14, 2021 Hello! Just a quick fyi that we won't be running columns this week (we're on semi-vacation at the moment). More tomorrow.:) Top News Investors are

Three brand new business opportunities! ⭐

Monday, June 14, 2021

Software Ideas Premium Newsletter Preview | Issue #52 View this email in your browser Opportunity #1: Programmatic advertising for small newsletters. Paved is a platform that helps advertisers launch,

Email Marketing Checklists

Monday, June 14, 2021

Our weekly round-up, top posts, upcoming events, and more! GrowthHackers Weekly Top Posts Welcome to the 1159 people who have joined our Newsletter since last Monday! Here is what you need to know from

Growth Bite: Put a feedback form on every screen of your product to boost retention

Monday, June 14, 2021

We all know how valuable customer feedback can be. Get more of it by offering users a simple way to make suggestions on every screen or page of your tool. Jason Forrest of [Rigbooks](https://www.

Running YouTube Ads without Going Broke 💸

Monday, June 14, 2021

Free Training Magazine PODCAST Articles Hey, Are you super pumped to start running YouTube ads but worried about costs? Well, here's some exciting news for you. We've done the research, and we

Startup Digest Reading List - Credit Scores, Market Infrastructure, and the Best Projects

Monday, June 14, 2021

A HUGE congratulations to Solana including Anatoly, Raj, and the whole team - on their milestone of $ Startup Digest Reading List Issue #159 — View online — Suggest a link Startup Digest Reading List A

MicroConf Mastermind Applications are Now Open

Monday, June 14, 2021

MicroConf Matching Cost: $100 ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌

Get paid for your feedback, join Wynter’s research panel

Monday, June 14, 2021

Sponsored Get paid for your feedback, join Wynter's research panel Want to give back to the community while having a low-key side hustle to fund your habit? Wynter is looking for people to join its

Today's startups in 10 words

Monday, June 14, 2021

Do you want to reach 1500+ developers, designers, investors, and entrepreneurs & support 10words? Check out our new sponsorship options. 10words.io Discover new apps and startups in 10 words or

Today's Digest: Submitted my Temporary Email extension to the Chrome Store 🎉

Monday, June 14, 2021

Your Indie Hackers community digest for June 14th ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌