Insight Innovation Ventures recently published Decision as a Service: The Architecture That Changes Everything, naming a shift the insights industry has been building toward without saying it. That shift? Selling the decision itself, tested and confidence-scored, instead of the report that used to sit in front of it. I think we can all agree with the thesis, but the question I keep coming back to isn’t whether the architecture is real, because obviously it is. The question is who can build it credibly. And that’s a much shorter list than the framing suggests.

Decision-as-a-Service isn’t a technology problem, ultimately. There are enough players in the space to prove that it goes beyond tech. The challenge is that it’s really a three-legged problem, and it’s rare to find anyone standing on all three legs at once.

The three legs

Data. I’m talking about unique data that hasn’t been licensed or scraped, real vs. synthetic. The data must be from a proprietary, human-verified panel, regularly refreshed, and detailed enough to calibrate a model against. It can’t just be AI gobbledygook.

Technology. The stack itself can’t be vaporware. It has to be smart and powerful enough to turn that data into something that operates effectively and continuously. It has to operate on top of a model that’s trained on the data, along with a decision layer that turns an audience definition into a strategy or a brief without a human needing to re-key it. Any integrations have to be able to activate the decision natively instead of handing it off to be rebuilt somewhere else.

Standing trust within the relevant ecosystem. Now this … this is the leg that’s overlooked in most architecture diagrams, but it’s really the one you can’t take a shortcut with. Getting this leg right requires both fluency and trust earned with the teams that will act on the decision, including the marketer setting the strategy, the agency planner building the brief, and the publisher sales team promoting the audience.

Look at IIV’s own cast of characters through this lens, and you can pick out the gaps. HelloTwin has real technology for decisioning and, by their own account, no proprietary body of data under it. They’re borrowing the first leg. Expert networks like GLG or AlphaSense have the third leg with genuine expertise and trust, but as IIV points out, they don’t compound because they don’t have a data asset or technology, and each engagement is a one-off. NIQ Cadence has data and technology, genuinely at scale, but its trust is with CPG retail, built with brand managers buying scanner data. It isn’t standing inside the agency-planner-and-publisher ecosystem where creative and media decisions are made and defended.

So we frequently see two legs, but rarely all three, because the third one can’t be bought or accelerated. It has to exist before the architecture needs it.

Our platform had two legs solidly in place before it even had a name.

• Big Village is the data leg. With a 90-year history in research, 100,000+ verified U.S. respondents, and over 260 variables per person (refreshed regularly), it incorporates synthetic data, but it isn’t a synthetic panel or a generic guess at a consumer persona.
• The platform is the technology leg. Models are fine-tuned against that unique panel, not a publicly available LLM. It’s a decision layer that enables an audience definition to inform a brand strategy, or to become a campaign brief or media plan without ever leaving the platform. From there, direct integrations into media endpoints ensure that the targeting defined in research is the targeting that runs across the strategy.
• We already had the third leg ready to go. Deep Focus is a successful agency, and they’ve been stress-testing this discipline in the market, against live client briefs, and under real competitive pressure. BrightStream is ad-tech and SSP infrastructure, and it closes the loop from intelligence to media. It stands within the agency-and-publisher ecosystem, and that position drives real results: an all-time single-day reservation record broken five times in one campaign for a travel and tourism client, and a 31% lift in high-value visitation for another that needed to stop chasing volume. This is more than a claim that our system works. It’s evidence.

A plain answer to the closing question

IIV ends their piece asking who builds full-stack Decision-as-a-Service for the market the enterprise players don’t reach, and who earns the right to be trusted with it. Our answer: Very few organizations have even two of these three legs under one roof, and even fewer have all three without having to borrow, buy, or fake one of them. We count ourselves among the few.

The platform launches in a few weeks, and it’s the clearest expression of a very old idea inside a new architecture: A company doesn’t get to sell the decision until it has earned the standing to be trusted with it.