Enterprise-grade data, analytics and media intelligence for brands and agencies across Australia.
Enterprise-grade data and media intelligence — without the enterprise agency around it.
BX3 Intelligence is led by Jeffrey Barnes, who spent 24 years building data, analytics and media science capability inside global networks including WPP, IPG and Omnicom — directing measurement and media strategy for some of the world's most recognisable brands, in the United States and Australia, with $2B+ in annual advertising investment under analytical management. That career covered the work most consultancies only advise on: building the models, the measurement frameworks and the teams, then standing behind the results in the boardroom.
We work principal-led: the person you meet on day one is the person doing the work. When an engagement calls for depth outside our core — creative technology, specialist engineering, regional market expertise — we bring in proven specialists from our network, deliberately and transparently. No bench, no pyramid, no junior team learning on your budget.
That experience spans 12+ industries — financial services, streaming and entertainment, retail, mobility, automotive, travel, telecommunications, healthcare and the public sector — which means pattern recognition most single-vertical consultancies can't offer.
When a consumer asks an AI assistant "what's the best option for X," the assistant answers with three or four brands. There is no page two. There is no retargeting the people who saw a competitor's name instead of yours. The recommendation simply happens, and you were either in it or you weren't.
Most brands have no idea which side of that line they're on.
The instinct is to treat this as an SEO problem with a new interface. It isn't. Search rankings respond to ongoing optimisation within weeks. AI assistants synthesise from years of accumulated public text — articles, reviews, forums, comparison pages — and that corpus is heaviest wherever a brand has been famous longest. A brand's AI visibility is something like a moving average of its last decade of reputation, weighted in ways nobody fully controls. You cannot buy your way into the answer, and you cannot fix it in a quarter.
That structure makes a prediction, and it's an uncomfortable one for any established brand: strong presence in the categories you dominated historically, and thin air in the adjacent, larger categories your growth strategy depends on. The assistants have learned your past, not your strategy. And because the topline mentions look healthy, nobody notices the shape of the gap until the growth categories underdeliver.
What you can do is measure it, and most organisations haven't even started. The method is not exotic: a few hundred realistic buying prompts, phrased the way customers actually phrase them, run repeatedly across the major assistants and scored for whether and how your brand appears. Three questions that measurement should answer for any brand:
Where do you actually show up? Not in the prompts your team would write — in the prompts your customers would. The phrasing a category manager uses and the phrasing a 34-year-old comparing options at 10pm uses produce different answers.
What's the gap between your visibility and your strategy? Strong presence in declining categories and absence in growth categories is the most dangerous pattern, because the topline number looks fine.
What's feeding the machine? The corpus AI assistants draw on is, in large part, your earned media, your review footprint, and the comparison content third parties write about you. That's a measurable, influenceable surface — but only if someone owns it.
AI visibility is becoming to the late 2020s what search visibility was to the late 2000s: a layer of demand that forms before any media impression is served. The brands measuring it now will set the benchmarks everyone else gets measured against.
Channel Economics · May 2026Most marketing organisations measure paid and owned channels in different rooms, with different teams, different dashboards and different definitions of success. Paid media gets judged on cost per lead. Email and owned audiences get judged on open rates. Nobody ever lines them up against the same yardstick — which is convenient, because the comparison is uncomfortable.
There's a simple discipline that exposes the problem: business value per thousand leads, computed identically on both sides. Same filters, same window, same definition of a qualified outcome. Few organisations have ever produced that number on a consistent basis — and when they do, owned audiences almost always carry far more downstream value per lead than any dashboard implies. The interesting question isn't whether there's a gap. It's where the gap lives.
The instinctive explanation is audience quality — "of course your email list converts better, they already know you." True, but incomplete, and the decomposition matters. Value per lead is only ever two things multiplied together: the rate at which leads become customers, and what each customer is worth once converted. Those two components point to completely different actions, so before drawing any lesson you have to know which one is carrying the gap.
More often than not, it isn't the one people assume. Converted customers tend to be worth roughly the same whichever door they arrived through. What differs is conversion mechanics — the proportion of leads that ever become customers at all. The owned journey has fewer steps, warmer context and better timing, so a far higher share of its leads complete the trip.
That distinction changes what you do about it. If owned customers were simply worth more, the lesson would be "owned reaches better people," and the action would be demographic targeting in paid. Because the gap is mechanical, the lessons are operational: the owned journey's properties — fewer steps, warmer context, better timing — are things you can study, and partially import into paid journeys.
Three implications we'd generalise to almost any acquisition-driven business:
First, measure both sides with one ruler before you set next year's budget. If you've never computed business value per thousand leads on a consistent basis across paid and owned, your channel mix is being set by whichever team presents last.
Second, treat the owned conversion journey as a design spec, not a benchmark. The question isn't "why does email convert better" — it's "which specific properties of the owned journey can paid traffic inherit?"
Third, fund list growth like media. If a lead on your owned base is worth a multiple of a paid lead over time, then the activity that grows that base deserves investment cases of the same rigour as your media plan — not the leftovers.
None of this argues for cutting paid media. Paid is how the owned base gets built in the first place. It argues for measuring the system as a system — because the moment you do, the budget conversation changes.
Measurement · April 2026A number drops year on year and the organisation does what organisations do: someone asks what went wrong, the channel team prepares a defence, and budgets brace for the consequences. The diagnosis is usually written before the analysis starts — performance got worse.
Take the most common version of the story: a portfolio's headline lead-to-customer conversion rate declines meaningfully year on year. The intuitive reading is rate softening — the same campaigns converting worse than they used to. It's also, very often, wrong.
Before accepting it, decompose the number. Any aggregate rate change can be split into two components: the rate effect (each segment's own conversion rate moving) and the mix effect (the blend of segments shifting toward lower- or higher-converting parts of the portfolio, even if no individual segment changed at all). It's a standard technique in finance — rate/volume analysis — and it's chronically underused in marketing.
Run the split and a different story frequently emerges: most of the decline is mix. The portfolio has shifted toward segments and campaign types that always converted at lower rates — often as a deliberate strategic broadening into new audiences or products. Like for like, within campaigns, conversion has barely moved.
Same headline number. Completely different story. And critically, a completely different set of actions:
If a decline is rate-driven, the fixes live in execution — creative fatigue, bidding drift, landing-page decay, lead quality erosion. If it's mix-driven, the right conversation is strategic: is the broadened portfolio generating enough volume at its lower rate to justify the blend? That's a question about intent and economics, not about whether the marketing team did their job.
The general lesson is bigger than any one portfolio. Aggregate metrics are composites, and composites lie by omission. A few rules we apply by default:
Never diagnose an aggregate. Any KPI that pools segments — conversion rate, average order value, cost per acquisition, retention — should be decomposed into rate and mix before anyone is asked to explain it.
Treat mix shifts as decisions, not failures. If the mix moved because strategy moved, the metric decline is the cost of the strategy, and the question is whether it was priced in. Punishing teams for it teaches them to stop broadening.
Build the decomposition into reporting, not post-mortems. If rate/mix splits only appear when someone's defending a bad number, they'll always arrive too late and look like excuses. Standing decomposition makes the conversation honest in both directions — it also catches the quarter where a flattering headline number is hiding genuine rate decay underneath.
The cheapest analytical capability most marketing organisations are missing isn't a new platform or a bigger model. It's the discipline of asking "what is this number made of?" before asking "who is responsible for it?"