Executive Summary

Most digital organisations believe they are data led and are investing in insight.

In reality, they are paying a hidden tax on their own curiosity. 

As the industry rushes to introduce AI to every platform, your data budget is subsidising an AI arms race you didn't sign up for. Platforms are passing on skyrocketing LLM costs through opaque pricing models, charging you for "innovations" that add zero value to your bottom line. Simultaneously, basic commercial questions have been moved behind a "pay-to-play" wall. Between premium credits and data-base costs, you are now paying, more than once, for data you used to access for free. 

This isn't just a cost issue; it’s a systemic threat to velocity. When every analytical "what if" carries a transaction cost, teams stop exploring and your commercial visibility become blurred. 

To reclaim your commercial reality, you don't need more software. You need a forensic measurement framework that strips away the noise. 

This article explores how organisations can stop spending on data and start investing in truth.

The Challenge

Many digital organisations appreciate the importance and the value of data. 

In everything from paid marketing to conversion analytics, data is the golden thread that ties together decision-making for organisations of all sizes, from founder-run start-ups to global multi-billion-dollar retailers.

But data, like everything, is getting expensive. 

As the market grows and larger providers shift their focus to AI, it is becoming harder for decision-makers to see through the haze, and more expensive for them to try. 

Rather than making the picture clearer, adding more data solutions makes it harder to see what’s real - when every platform has its own quirks and nuances that affect reporting adding more data, more solutions, and more cost, will not make your data clearer. It will simply increase your software costs.

Paying Twice

Imagine you wanted to answer a specific question of customer engagement: “What proportion of users who land on our homepage via Google Organic Search click on the ‘Most Popular’ rail and subsequently add a product to the cart?”. A simple enough question, but not something immediately available out-of-the-box in any data platform. 

In years past, you would answer this by configuring a segment in Google Analytics, producing a data table, and perhaps extracting it via the API. You may also leverage sampling mitigation if you had significant traffic volumes (but this didn’t cost you anything). This process was cheap, quick, and scalable; you could build expansive customer insight frameworks, almost without any cost.

Today, answering that same question has become a “pay-to-play" exercise. 

Mitigating data sampling now requires premium credits or the expensive GA4 360 platform. The reporting API no longer accesses segments, forcing you to extract data to BigQuery where every nuanced data query you run when every query, no matter how small, adds a cost to your bill. Third-party solutions that you could once rely on have been acquired, franchised, and now cost more.

The real danger here isn't just the invoice; it is the erosion of velocity.

When every analytical question carries a micro-transaction or a credit cost, teams stop exploring, they stop asking "what if" because the budget won't allow it. 

Today, you aren't just paying a premium to access the insights you used to have for free - you are paying a tax on the very curiosity you must encourage.

Subsidising the AI arms race

Frankly, one of the major causes of this cost increase is the intense focus on rolling AI into everything.

As data platforms bundle AI into their service offering, their overheads skyrocket. Remember: these platforms rarely own their own AI. They are simply connecting to ChatGPT, Claude, or Gemini. This costs them tokens, and they are passing that cost directly onto you through their pricing models, which you’ll note never tell you the cost of the AI features, that you cannot opt out.

Even if you never touch these AI features, you are still paying for those tokens. 

And as AI models feel the pressure to show profitability, the cost-per-token will increase, raising the floor for every platform you use. You are paying more to measure the same thing, without receiving a single byte of additional data and not a single £ of additional value.

The cost of failing to act

Without addressing this challenge, organisations are going to find themselves battling several systemic issues:

Eroding Margins: It is going to become significantly more expensive to obtain reliable data and insight.

Stifled Innovation: Budgets will be stretched, forcing decision-makers to sacrifice value-adding solutions because the "data tax" has eaten the budget.

Analytical Noise: More low-cost, low-quality data solutions will enter the market, create more confusion and further slowing decision-making.

If you cannot get control of your investments in data, no amount of software or AI will overcome the fundamental challenge. Your commercial visibility will blur, and you will lose the ability to respond to the customer, who will simply go elsewhere.

What you can do about it

Fortunately, digital analytics, everything from marketing to returns, can be distilled into four core categories:

See: What does the customer see (pages, ads, products, scroll depth)

Click: What do they interact with (ad clicks, product images)

Do: What behaviours do they complete (purchases, sign-ups)

Spend: How much do they spend, and what do you spend to get them

The four data categories roll up into the most important one – that if you can answer you can unlock digital sales:

Fail – Why do customers fail to buy

To take back control of your data and your investments, you first need a robust digital measurement framework; a forensic list of the metrics, dimensions and segments that actually describe your commercial reality. Once this is in place, you can map the data you need against the data you have.

In most cases, this will reveal that you already have the data you need, or that gaps can be filled with strategic tagging via Google Tag Manager rather than implementation of another data technology. You no longer need expensive heatmapping or "behavioural" suites when you can achieve the same results through a smarter measurement strategy.

Not only does this allow you to reclaim budget that no longer has a justification, but you will obtain a clearer view of your commercial reality. This makes everything easier, from measuring performance to prioritisation and decision making, you will no longer spend hours battling through dashboards and reports but can quickly get to the truth and act accordingly.

Conclusion

The promise of the modern data stack was clarity, yet for many digital retailers, the reality is a nothing more than bloated line item on the P&L that offers less visibility than a spreadsheet did five years ago. 

By allowing your data strategy to be dictated by platform "innovations" - most of which are simply ways to pass on the costs of an AI arms race - you aren't just losing money; you are losing the ability to lead with confidence.

Effectively scaling sales in digital channels comes not from simply increasing spend, but through a forensic understanding of the customer journey, not a suite of AI-powered "black box" tools that charge you for insights you already own. 

Reclaiming your commercial reality starts by stripping back the noise. When you stop paying for the privilege of looking at your own data and start focusing on a lean, strategic measurement framework, you stop spending on data and start investing in the only thing that matters: the truth of how your customers interact with your brand.

Your data architecture should not feel like a sunk cost but the key to your commercial success, by focusing on what you truly need rather than that you’re being told you need means you can measure once, pay once, and use the savings to drive the growth your data was supposed to find in the first place.