The Digital Insider · October 2026
The expensive mistake is answering the wrong question
We published four pieces this month, for four very different readers: a Drupal team choosing a front end, a deal team signing an acquisition, and two sports organizations, one thinking about AI and one about memberships.
Reading them back together, we noticed they all make the same argument. In each case, the team thinks it's making one decision when it's actually making two. The cost comes from answering one while thinking about the other.
Here are all four. If you only have two minutes, read the bold lines. We won't be offended. We'll just be a little sad.
"Are we going headless?" is two questions. Find your box.
If you've sat in a Drupal planning meeting this year, you've probably heard someone say "we should go headless" with the confidence of a person ordering the special without asking what it is.
The curiosity is fair. In WP Engine's State of Headless 2024 survey, 73% of businesses said they already use a headless architecture, and 65% named budget as a barrier. (WP Engine sells hosting, so read the enthusiasm accordingly. The budget number we believe completely.)
Here's what years of untangling these projects has taught us: "headless" usually stands in for two separate questions. Answer them one at a time and you get four boxes, all legitimate.
Or start from who you are:
Our one piece of advice for every box: build Code Components, not Twig templates.
Components come with you if you change boxes later. Templates stay behind, like the sofa that never fits through the new door. And one old line aged badly this month: "headless means two vendors." Acquia announced its alliance with Vercel on 28 September, so two systems no longer have to mean two relationships.
Read: One pantry, four ways to cook →Your deal model counts revenue, headcount and synergies. It may not count the estate.
Here's a sobering pair of numbers.
Put those side by side and you'd expect the digital estate to be the first thing a deal team examines. In our experience, it's usually the last, and someone who wasn't in the deal ends up paying for it after close.
An acquisition can buy you a second set of everything: another website, another storefront, another set of customer records, another set of reporting definitions. Not all of it reaches the letter of intent. All of it needs an owner on day one. You don't need a technologist in the room to start asking. Tick off the ones your last deal actually covered:
Be honest. Nobody's watching.
Why number four goes first
A failed CRM rollout leaves behind something no new software can fix. In one multi-brand group we worked with, the sales team had quietly gone back to spreadsheets because nobody trusted the system anymore. That's a trust problem, and it was there to find before anyone signed.
Number six has a sting in its tail. If one business counts a deal at proposal and the other at signature, your own buyer's diligence will find it one day, probably at the worst possible moment.
Read: The cost diligence never checks →What a race-day assistant taught us
If your inbox has filled up with "agentic AI" pitches lately, you're not imagining it. Gartner expects over 40% of agentic AI projects to be cancelled by the end of 2027, citing rising costs, unclear business value and weak risk controls. It also reckons only about 130 of the thousands of vendors claiming agentic AI are the real thing.
So we'd rather tell you about one small agent that worked, and why. For a participation-sports organization, we built a pilot assistant that answers athletes during race week, using approved race content only. Go on, be an athlete for a minute:
An illustrative demo, based on the pilot's rules. It doesn't hold real race data.
Three things surprised us:
Our take: a narrow assistant that's right beats a broad one that's sometimes wrong.
Start with questions that arrive often and are cheap to get wrong, like race-week logistics for a single event. Measure accuracy and handoff rates alongside time saved, because in this business a wrong answer costs more than a slow one.
Read: Start with guardrails →Most participants don't come back to the same event
Now hold those numbers next to one of the most quoted findings in business. Research by Fred Reichheld at Bain found that raising customer retention by 5% can increase profits by 25% to 95%. For an event organizer, that gap is the whole opportunity.
Pro sports has noticed. Clubs are swapping season tickets for monthly passes and tiered memberships. Participation sports still mostly sells one entry at a time and hopes people come back next year. We think memberships fit participants even better than fans. A fan's membership buys access to watch. A participant's can buy progress.
We see four models emerging. Tap a card to see who it's for:
Two of us, four cities, one very full inbox of notes
VP Sales & Partnerships
Director of Sales & Partnerships
Piyush spent the turn of the month in Rotterdam, Manchester and Edinburgh. Then Sachin took the passport to Philadelphia, to a room full of people who buy companies for a living. Pick a city:
If one of these questions is sitting on your desk right now, reply and tell us which. A person reads every reply, and we're happy to think it through with you, whether or not we end up working together.
From the team at Axelerant