On holiday packages and the ad tech stack you didn’t book

It’s that time of year again … conference and holiday planning season. Ad tech folks crossing oceans for pleasure, for business or both. And while everyone is busy planning the next trip, nobody really argues about the right way to do it. One of the few remaining “live and let live” realms.

Option A: Some people book everything through one operator. Flights, hotels, transfers, excursions and the lot.

Option B: Others pick their own flight and hotel, but they take the hotel tour bus to the ruins because it’s there and it’s easy.

Option C: Some plan every element themselves. Different vendors for everything, flight, Airbnb, rental car, each chosen on its own merits, offline maps downloaded, just in case something breaks.

All three are valid choices. The debate about which is correct is, mercifully, not a thing.

The programmatic industry has spent the better part of a decade trying to have that debate anyway. Walled gardens: the package holiday. All-inclusive, end to end. One operator, one invoice, one set of numbers that reliably show the campaign working. Google, Meta, Amazon, different companies, same promise: Book everything through us; here are your results. Done and dusted. The attribution is convenient; the reporting is clean. The hotel tour leaves at 9, and you’ll be back by lunch.

The interoperable open approach sits with platforms like The Trade Desk. A stack that delivers the benefits of independence (transparency, control, third-party optionality) without requiring buyers to architect everything from scratch.The hotel won’t stop you from leaving for dinner. Some guests don’t bother, but they could if they wanted. Both are legitimate uses of the platform, and that’s the point. The choice lives with the buyer, not the vendor.

And then there’s the maximum customization stack. DSP, measurement, brand safety, curation, all handpicked, all separately contracted, all requiring someone to own each relationship and reconcile the outputs when they disagree. Which they will. This is the person who researched every restaurant, hired a local guide and still checked Tripadvisor at the table. They probably got more out of the trip, and not by accident. But they also put far more effort into planning it.

The ad tech industry, as a whole, has developed a habit of framing these options as ideology, a controversial topic that’s primarily discussed on stages, in trade press, and across LinkedIn. Open web versus walled gardens. Independence versus convenience. Principled practitioners versus lazy ones. It’s none of those things and all of them at the same time. It’s a product choice, made under constraints that often have very little to do with what anyone actually believes.

Take independent measurement. In practice, it means a vendor decision, a contract, implementation and then the genuinely awkward part: a set of numbers that may not match the platform’s reporting. Someone has to explain that to a client or a CFO who doesn’t want a nuanced conversation about methodology. They want the campaign to have worked. Vendor measurement, in this context, isn’t just easier; it’s organizationally safer. The risk of a bad number is diffused. Nobody is accountable for an outcome the platform itself is measuring. This is incentive design at its purest.

Team expertise compounds the problem. Running a genuinely independent stack requires people who know what good looks like, who can evaluate a measurement vendor, stress-test a curation partner’s claims and spot when a brand safety tool is earning its fee versus just adding a line to the invoice. That knowledge does exist. It’s expensive, unevenly distributed and has a habit of leaving organizations at exactly the moment a restructure happens.

But set contracts and head count aside. Do media teams even have the bandwidth, knowledge and patience to choose Option C? Not in theory, in practice, on a Tuesday afternoon, with three campaigns live and a client call at 4. For most, the honest answer is no. And that was true before AI entered the picture. What AI does is make the end-to-end package (ideally) better, faster, cheaper and increasingly capable of doing things that once required either deep expertise or significant budget. The walled garden isn’t standing still. It’s improving at a rate that the independent ecosystem, fragmented by definition, will struggle to match on convenience alone. The holiday package doesn’t require anyone to speak the language.

Company culture determines whether this is even a live question. Organizations with a genuine testing-and-learning culture, where interrogating results is normal and better methodology is progress, tend to move toward independent solutions over time. Organizations with a delivery culture, where the priority is hitting numbers and avoiding new conversations (read: problems), tend not to. Let’s just say the second type is far more common.

And then there are the structural incentives the industry mostly discusses in private. Holding group volume commitments. Preferred vendor agreements set several layers above the person writing the media plan. Joint business plan (JBP), annual incentive plans (AIPs), joint technology initiatives (JTIs), a variety of plans and acronyms describing the exact same thing. Carrying the accumulated weight of last year’s contracts, rolled over and over again. Revisiting them requires a kind of organizational will that is genuinely scarce.

The buyer who knows that an independent tool would give a more accurate read often can’t act on it. The decision was made before they were in the room and will be made the same way next year.

And no, this is not a small part of the market. It describes a lot of it.

None of this is an argument for any one approach. The fully independent stack isn’t right for every organization, and pretending otherwise is its own kind of ideology. What it is an argument for is the choice being a real one, made deliberately, by someone with the information and authority to make it, with some honest accounting of what is driving it.

Some advertisers will always choose the package. Some will always want the offline maps. Most will do different things at different times, for reasons that are rarely purely philosophical.

What makes this existential is that the open ecosystem’s pitch has always relied on an implicit assumption that buyers who care enough will do the work. That the delta in quality justifies the delta in effort. That Option C, for the right organization with the right team, is worth it.

But was maximum customization really ever the point? No, not really. The point is maximum openness. Open as in transparent. Open as in interoperable. Open as in finally freeing up ad ops calendars. More control. Cleaner workflows. Fewer reconciliation headaches. Basically, the benefits of independence without the operational burden that has made them feel out of reach.

This op-ed represents the views and opinions of the author and not of The Current, a division of The Trade Desk, or The Trade Desk. The appearance of the op-ed on The Current does not constitute an endorsement by The Current or The Trade Desk.