Today Is The Perfect Day To Fail

Actually, yesterday was better. But today will do just fine.

So you must try to fail today. Please.

This sounds absurd until you understand the economic moment we're living through. AI tools are heavily subsidised by platform titans burning billions to gain market share. Meta, Google, OpenAI, and Anthropic are competing to own the AI infrastructure layer, which means they're pricing tools well below their actual cost to deliver.

This creates a narrow window where experimentation is artificially cheap. Where building genuinely better capabilities and not just faster or cheaper versions of what you already do, costs less than it will ever cost again.

And yet, according to Gartner's 2025 CMO Spend Survey released last week, most marketing teams are using this subsidised moment to optimise for exactly the wrong things.

For the past nine months at Simbioniq, We've deliberately pursued failure. We chose to build psychologically complete simulated humans rather than faster surveys or cheaper focus groups.

That choice was only economically rational because failure is currently subsidised, and our resulting technology now has a patent pending.

When AI pricing normalises, and it must, to cover the $600 billion infrastructure investment required in the space, our competitors will face dramatically higher costs when they finally attempt to build better rather than just faster and cheaper.

The window is closing. So if you're a CMO and your teams are only reporting time savings and cost reductions, you should be genuinely concerned. Because the absence of failures means the absence of attempts at better. And better is the only thing that survives when the subsidy ends.

The major AI platforms are in a classic infrastructure land grab. OpenAI raised $6.6 billion in October 2024 at a $157 billion valuation while reportedly losing $5 billion annually on ChatGPT operations. Anthropic raised $7.3 billion across 2024. Google is integrating AI across its entire product stack despite massive compute costs. Meta is giving away Llama models to prevent OpenAI's API from becoming the standard interface.

Amazon didn't become the everything store by pricing profitably during growth phase. Netflix didn't achieve streaming dominance through sustainable pricing. Platform economics require subsidising usage during the land grab, then monetising after achieving scale.

AI platforms are following the identical playbook. OpenAI needs Microsoft to fund compute costs. Anthropic needs Amazon and Google backing. Meta needs users locked into its AI tools before competitors do.

What you cannot do is assume current pricing represents sustainable economics. When platforms need to show return, pricing will rise to match utility value. The current arbitrage opportunity is a temporary market inefficiency created by platform competition.

Gartner's survey reveals exactly how most CMOs are using this subsidised moment. Among organisations deploying AI:

The Easy Wins Look Like:

Do you recognise these wins in the AI progress that you have been celebrating?

Every single one of these capabilities is available off-the-shelf. They are table stakes. Every major platform offers these tools. Every competitor in your market can deploy identical capabilities in weeks.

These are easy precisely because they're subsidised and productised. They require minimal organisational change, deliver measurable efficiency metrics quickly, and satisfy board pressure to "show AI progress."

But easy wins available to everyone create zero competitive advantage.

Simple economics: When AI platforms need to show a return on their investments, pricing rises to match utility value.

If an AI tool genuinely delivers 10x productivity improvement, sustainable pricing captures a significant portion of that value. Current pricing doesn't reflect this because platforms compete for market share, not profit.

Teams that cut headcount based on subsidised pricing face a crisis: They eliminated human capabilities assuming AI cost savings were permanent. When AI pricing rises to sustainable levels, their "efficiency gains" evaporate while their human capabilities are gone.

They're locking in structural changes based on market inefficiency that platforms will eliminate once they've achieved sufficient scale.

Building better towards genuinely enhanced capabilities, not just faster execution of existing processes, requires something easy wins never demand: failure.

Better means attempting capabilities that don't exist yet. Approaches that might not work. Team changes that could fail. Strategic bets that require multiple iterations before delivering value.

Better looks like:

None of these arrive through deployment of productised tools. All require experimentation, failure, learning, and iteration.

The teams that will dominate in five years are failing right now. They're attempting better while subsidies make failure affordable. They're building proprietary capabilities while everyone else deploys commodity tools.

Here are the camps, according to cluster based on the Gartner research:

The Cost Cutters (78% of CMOs):

The Status Quo Optimisers (17% of CMOs):

The Capability Builders (5% of CMOs):

Where are you? Be honest now. Then go out and fail today, please.

The best time to fail was yesterday, when subsidies were even larger and pricing was even cheaper. We cannot recover that moment.

Every quarter you wait, failure becomes more expensive. Every month you spend optimising for faster/cheaper, you fall further behind organisations building better.

The question isn't whether to fail. It's whether to fail now while it's cheap, or fail later when it's expensive. Because building better requires failure, the only variable is the cost.

So I'll say it again: You must try to fail today. Please.

Not recklessly. Not without learning. Not without strategic intention.

But deliberately, proudly, systematically. Attempt capabilities your competitors aren't building. Test approaches that might not work. Build proprietary advantages that require iteration.

And when your teams report failures alongside efficiencies, celebrate them. Because failures during a subsidy window are investments in capabilities that matter when subsidies end.

The teams that dominate in five years won't be the ones that cut costs fastest in 2025. They'll be the ones that failed most productively while failure was cheap.

So fail today. While you still can afford it.

Brandflow is written by Justin Billingsley, who has spent his career on all three sides of the industry's table: senior client, global agency leader, technology founder. First published 12 January 2026 in the Brandflow newsletter on LinkedIn.