Marketers Are Optimizing Away Their Own Luck
Serendipity is a martech skill. Most teams optimize it out, especially in the operational half of the stack nobody explores.
This newsletter is a remix of a previous one that tackled serendipity and martech (no, not an oxymoron, they can actually go together!).
I ended up so interested in this unlikely pairing that I ended up adding it to my own book: “Strategic Marketing Skills: How to build technology-driven expertise that delivers business value”
Below, I also give a personal example of serendipity inside a martech stack I managed!
Read it and let me know if you agree with this example or what you would consider serendipitous when it comes to your stack!
A couple of years ago I was in the audience at our email platform’s event when they announced a new feature, and it clicked in my mind right away. I had to idea this was coming up!
Our email program ran on two (main) dependencies. One team built the HTML templates. Another managed the data and analytics that fed them. Every step up in cadence or personalization meant more load on one of those teams. The feature on stage could automate the customization of email templates straight from customer data, which meant we could scale the program without leaning harder on either team (meaning, without increasing the cost of building these emails).
I did not wait for a planning cycle. I raised my hand on the spot to join the beta as one of the first testers. When I brought it back to the teams who would co-own it with me, they moved fast. We ran it as a test, and it has since become a core part of how email programs were run.
(small promo to my Brazilian friends 😉)
(end of promo)
Here is what it actually does: the entire email template, every component, is generated from a simple spreadsheet and customized with customer data. Any business stakeholder can keep that automation current (since all it takes is updating an spreadsheet). The only code involved is the double curly-bracket variables in a dynamic template, and that is exactly the kind of thing AI is already pretty good at now.
The best thing my team shipped that year started as a feature I happened to be in the room for, and a decision to raise my hand before anyone asked me to.
I have thought about that sequence a lot, because almost nothing about how we usually work in martech is built to let this type of thing happen.
In The Serendipity Mindset, which sits on the Experimental Marketer Library shelf, Christian Busch defines serendipity as chance meeting human action. You take part in it. It needs three things at once. Agency, so you are actively engaged. Surprise, so the outcome is unexpected. And value, so you recognize the find as worth something. My little email story had all three. The announcement was the surprise. Raising my hand was the agency. The teams reacting and scaling it was the value. Busch’s book earns its place on a martech shelf for a simple reason. It is really about connecting the dots between technology and process improvement.
You are reading The Experimental Marketer Newsletter.
What you get: every week, one martech or AI concept broken down
with real examples, step-by-step walkthroughs, and the career case for why it matters.
Created by a Marketer, for Marketers.
Most martech work is built to remove exactly the conditions that story needed.
Marketers get laser focused on their own platforms and their own way of working. That focus is a strength but, at the same time, it is also a wall. When every hour goes to running the system as designed, there is no space for the accidental detour where a better idea can live. I wrote about this in my book's chapter on strategic discovery, because the opportunity is larger than most teams realize.
Look at what the focus can hide. A marketing director runs a routine budget review and finds her company pays for 47 separate martech tools, and that the team can account for only about 30 percent of the functionality across them. Three departments pay for the same data enrichment service, none aware of the others. The wasted budget is the obvious problem. The quieter one is everything those unopened features could have done, for example.
And the features most likely to stay unopened are the ones that can help increase efficiency and efficacy of the marketing operations, the workflow automations, the data connections, the small process shortcuts. Many times, these rarely gets the attention when compared to customer-facing activities. My email automation lived there, in the operational half of the platform we could further explore for efficiency and efficacy. The gold, as the strategic-discovery work puts it, is buried in the ecosystem you already have, not in the next tool you buy. A new tool laid over the same process gaps just runs the same errors faster.
The designer and technologist John Maeda has a name for the useful version of this. He calls it the art of desirable accidents. The skill to improvise, he writes, is to go for it and realize a desirable accident. His point is that these accidents are not pure luck. They come from creating the conditions where productive accidents can happen, and then having the wisdom to cultivate the best ones that emerge. Remove the fear of error, and you open the door to the discoveries that lead to your most original work.
Scott Brinker, whom AdAge called the Godfather of Martech, also shared that idea with the marketing technology world for a reason. It describes exactly the environment most marketing teams do not build for themselves.
The catch is that a desirable accident only pays off when someone notices it. A stale automation, an odd result, a workaround that quietly beat the official process. The system does not flag it as a win. A person has to be paying attention and willing to treat it as a lesson rather than a defect.
There is a disciplined name for the useful version of this. Amy Edmondson, the Harvard professor who studies how organizations learn, calls it intelligent failure. It is the undesired result of a thoughtful experiment in new territory, the kind you had good reason to believe might work. Her argument is that these are the price of progress, worth welcoming rather than hiding.
Most systems are built to hide them instead. A stale automation, an odd result, a workaround that quietly beat the official process. The system logs it as a defect. Someone has to notice, and be willing to treat it as a lesson, before it turns into learning.
Two shifts make this easier to do on purpose. Marketers are starting to build their own internal tools. In Vibe Coding in Marketing, a 2026 report from chiefmartec and UserEvidence based on more than 300 marketing leaders, most teams using AI to generate code were pointing it at behind-the-scenes tools rather than customer-facing ones, with the biggest gains in the operational work that used to depend on engineering, like data integration, reporting, and workflow automation. That is the same shift that lets a business stakeholder keep my email automation current with a spreadsheet and a few curly brackets ;-). The cost of trying something has really dropped, which means the cost of a desirable accident has too.
None of this argues against control. Frans Riemersma describe marketing operations as two environments at once. A Factory that is stable, governed, and low error, running the campaigns the business depends on. And a Laboratory where the boundaries get tested and the accidents are allowed. The mistake is running only the Factory and calling the Laboratory a distraction.
That is the balance the job actually asks for, and it is really hard to find 9and even harder to keep!). Marketers have to stand in two places at once. Optimize relentlessly for what already exists, and hold space for curiosity and the new. Run the system, and keep poking at it. I think nobody resolves that tension permanently. The work is to keep both alive at the same time.
What you can do now
The practical move is smaller than a transformation program. Give yourself and your team explicit permission to explore the stack you already own, the internal operations features as much as the customer-facing ones, and to treat an accidental find as a real outcome rather than a distraction from the plan. Go to the event. Raise your hand. Open the part of the platform nobody has touched. Keep a Laboratory running next to your Factory.
Serendipity will never appear on the roadmap. It shows up for the marketers who left a door open for it, and then paid attention when something walked through.
End Notes
1. Christian Busch, The Serendipity Mindset: The Art and Science of Creating Good Luck (Riverhead Books, 2020). Source of the definition of serendipity as chance meeting human action, and the three conditions — agency, surprise, and value. Featured title in the Experimental Marketer Library.
2. Gartner Marketing Technology Survey. Marketers report using roughly a third (about 33%) of their martech stack’s capabilities, down from 42% in 2022 and 58% in 2020. Reported via MarTech, “Marketers are only using one third of their stack’s capability.” https://martech.org/marketers-are-only-using-one-third-of-their-stacks-capability/
3. The 47-tool budget-review scenario is drawn from the author’s own Strategic Marketing Skills (Kogan Page, 2026), Chapter 1, “Strategic Discovery.” It is an illustrative composite, not a named company.
4. John Maeda, “The Art of Desirable Accidents: Embracing Improvisation,” Medium. Source of the “desirable accidents” idea and the line, “the skill to improvise is to unhesitatingly realize a desirable accident.” https://johnmaeda.medium.com/the-art-of-desirable-accidents-embracing-improvisation-811b7ba456da
5. Scott Brinker, called the “Godfather of Martech” by AdAge, shared and amplified Maeda’s essay to the marketing community. https://www.linkedin.com/posts/sjbrinker_the-art-of-desirable-accidents-embracing-activity-7318992873574584320-vCJn
6. Amy C. Edmondson, The Right Kind of Wrong: The Science of Failing Well (Atria Books, 2023). Source of “intelligent failure” — the undesired result of a thoughtful experiment in new territory.
7. “Vibe Coding in Marketing,” chiefmartec (Scott Brinker) and UserEvidence, 2026. Survey of 302 marketing leaders. Cited finding: teams apply AI code generation mostly to behind-the-scenes and operations tools rather than customer-facing ones. https://userevidence.com/vibe-coding-marketing-report/ ·
8. Scott Brinker and Frans Riemersma, Martech for 2026 (2025). The report frames marketing operations as two environments, "The Factory" and "The Laboratory." Full report: https://content.martechday.com/martech-for-2026.pdf · Announcement: https://chiefmartec.com/2025/12/heres-your-copy-of-our-martech-for-2026-report-free-and-ungated/