Business Complexity Does Not Go Away with Mike Heilig of Binary Anvil
The pitch for moving to SaaS is that it takes work off your plate, that somebody else handles the hosting and the patching and the endless small maintenance that eats a team’s week, and Mike Heilig, general manager at Binary Anvil, grants every bit of that before taking issue with what merchants assume it buys them.
“They reduce complexity with hosting and patching and things like that,” he says. “But they shift the complexity somewhere else, because business complexity doesn’t go away.”
Binary Anvil is a web development agency and systems integrator that has been running for 18 years, built mostly on Adobe Commerce, and has since branched into other platforms and a product of its own.
The Complexity Moves
Heilig’s test is a question of fit.
“If you run a very simple business, maybe SaaS is the right fit for you,” he says. “If you have B2B requirements, complex catalogs, complex price lists, maybe SaaS isn’t the right fit for you.”
The failure he sees is upstream of the platform decision. Merchants shop for technology before they have written down what they want the business to do.
Merchants fixate on the latest trends and the latest technology, he says, when the useful move runs the other way. “Figure out what your business outcomes are that you want to achieve,” he says, “whether that’s improving your search or smoothing out your checkout process. Those are the things that matter.”
Put the other way around, the question is whether you are picking a platform that fits your business or reshaping your business to fit a platform.
What Decoupling Bought D&B Supply
Binary Anvil’s own product is Flux, a headless front end that attaches to an Adobe Commerce backend. D&B Supply, a large farm and ranch retailer, had a traditional Adobe Commerce build and wanted more fulfillment options than in-store pickup, better search, better performance, and room to grow.
Rebuilding the customer experience on Flux while keeping the Adobe Commerce backend produced a 49 percent improvement in site speed, a 67 percent increase in conversions, a 69 percent increase in organic traffic, and a 52 percent improvement in search rankings. Their annual upgrade cost dropped by roughly 57,000 dollars.
That last number is the one Heilig keeps coming back to, because it explains why Flux exists at all.
Where PWA Studio Stalled
Adobe’s own answer to a headless front end was PWA Studio, and it never got finished. Asked how Flux avoided the same fate, Heilig does not reach for a technical explanation.
“Well, stubbornness, I guess, is what I would say,” he says. “And maybe my military background made me want to finish it.”
The plan from here is to make Flux less specific rather than more. Instead of the front end reaching out to a backend and asking for a Magento product or a BigCommerce order, he wants it asking for the idea of an order, with the mapping handled in the admin or in middleware.
Upgrades Are a Front End Problem
Ask Heilig how he shrank the cost of an Adobe Commerce upgrade and the answer is architectural.
Highly customized front ends are the reason people choose Adobe Commerce in the first place, and they are also what makes every upgrade expensive, because each customization has to be dragged forward. Decouple the front end and the backend upgrade stops touching it.
“We’ve seen 70 to 80 percent reduction in cost of our upgrades for minor versions,” he says.
He also gives Adobe credit on a point that cuts against the SaaS pitch. Adobe has been careful with its API service contracts, and when a change is needed they tend to publish a new API and keep maintaining the old one. A custom headless build sitting on a SaaS platform carries the opposite risk, which is that the API underneath it gets deprecated on somebody else’s schedule.
Structured Data Before Agents
On AI mediated commerce, Heilig thinks the requirement is coming and is not here yet. What merchants can do now is unglamorous groundwork.
“Focus on making sure your structured data is in place,” he says. “Without that, AI doesn’t really know what it’s looking at.”
Machine readable documentation and clean APIs are the rest of it, so an agent knows what it is interacting with. Beyond that, he is wary of building for demand that has not arrived, and he borrows a framing from Roy Rubin about putting the cart before the horse, building functionality before there is a need for it.
Then there is the part nobody has solved.
“Are people really ready to hand their credit card information over to an AI agent?” he asks. “I’m not sure.”
The Community Built the Thoughtful Parts
Two of the features Magento users lean on hardest, multi source inventory and Page Builder, came out of the community rather than a vendor roadmap. Heilig thinks that shows in the work.
“Because of that they were more thoughtful in how they built it,” he says, and envisioned something that would help clients going forward, rather than an add on assembled in a hurry.