The Missing Middle (How Performance Marketing Lost its Way - Part 1)

Part 1 of 2: We draw on our founders' experiences to consider how Performance Marketing developed from earlier, simpler times, to the fast-moving and complex landscape of today. (Read Part 2 to find out how Performance Marketing found its way)

In 2005, Facebook—then still "The Facebook"—was running Google AdSense on its own platform. That detail might seem almost quaint today, but it perfectly captures how fundamentally performance marketing has transformed in less than two decades. We've watched this evolution first hand, from those early days of manual optimisation to today's AI-driven landscape. But somewhere along this journey, a critical disconnect developed that's holding back marketing teams worldwide—and understanding this gap is key to understanding where performance marketing needs to go next.

Those early days of performance marketing were marked by their simplicity—not because the work was easy, but because the relationships between action and outcome were crystal clear. Marketers would often have two screens arranged like a trader's desk: one showing campaign metrics in real-time, another displaying internal revenue data. They'd watch both simultaneously, making instant connections between spend and results. When performance dipped, you knew exactly why. When conversion rates dropped, you could trace it to a specific change. If a website checkout process broke for two hours, you saw it immediately in your numbers. There was no need for complex attribution models or cross-channel analysis; the cause-and-effect relationship was immediate and obvious.

The role of incrementality measurement was straightforward too. Unlike today's complex multi-channel environment, you could easily see if search was actually driving traffic or if your spend was delivering results. The numbers were right there in front of you. Your primary concern was simply whether the ROI made sense and if you should adjust your bidding strategy. The metrics weren't buried under layers of dashboards or hidden behind complex attribution models.

The challenges weren't about data complexity. They were purely about execution. Every decision about bids, budgets, and targeting was in our hands. We had to manually pick keywords, monitor their performance, adjust bids, and optimise budget allocations. There was far less automation than today, which meant more human intervention was needed to achieve optimal performance. But this manual control had an upside: we could react, adjust, and optimise in real-time because we had direct control and immediate visibility. When something wasn't working, we could fix it immediately. When something was working well, we could scale it up right away. This was performance marketing in its purest form: see a problem, fix a problem, measure the results—all within the same day, often within the same hour.

Phase 1: Channel Explosion

Around 2010, things began to change rapidly. The relatively straightforward world of search advertising exploded into a dizzying maze of channels, each with its own rules, metrics, and optimisation approaches. Search itself became more complex - what started as simple keyword bidding evolved into a sophisticated ecosystem of match types, quality scores, and ad extensions. And this was just the beginning.

Social media emerged as a major advertising force, with Facebook leading the way. Unlike search, where user intent was clear, social media brought new challenges of audience targeting and creative optimisation. You weren't just bidding on keywords anymore; you were targeting complex combinations of demographics, interests, and behaviours. The same customer might see your ad on Facebook, search and click through from a Google ad, and finally convert only after seeing a retargeting display ad - creating entirely new challenges in attribution and measurement.

Phase 2: The Rise of Programmatic Buying

Display advertising underwent its own revolution with the rise of programmatic buying. The old model of directly negotiating with publishers gave way to real-time bidding and complex audience targeting. Suddenly, marketers needed to understand concepts like viewability, brand safety, and fraud prevention. Each display network had its own set of targeting options, creative specifications, and performance metrics.

Mobile advertising emerged as its own distinct channel, with app install campaigns requiring entirely different approaches to measurement and optimisation. The introduction of app stores created new conversion points to track, while mobile web brought cross-device attribution challenges. What counted as a conversion when a user saw an ad on their phone but purchased on their laptop?

Phase 3: Multi-Channel Complexity

Affiliate marketing evolved from simple revenue sharing to complex networks with their own tracking, attribution, and payment systems. Video advertising exploded across platforms like YouTube, bringing new metrics like view-through rates and cost-per-completed-view. Even email marketing, one of the oldest digital channels, became more sophisticated with improved targeting and automation capabilities.

Each of these channels existed in its own silo, with its own reporting interface, its own metrics, and its own definition of success. A "conversion" in Google Ads meant something different than a "conversion" in Facebook. A "view" in YouTube wasn't the same as a "view" in a display campaign. Cost-per-acquisition calculations varied across platforms, making cross-channel comparison nearly impossible without significant data manipulation.

Phase 4: The Spread of Spreadsheets

This complexity gave birth to the era of the marketing spreadsheet. What started as simple campaign tracking evolved into sophisticated workbooks with multiple sheets, complex VLOOKUP formulas, and intricate pivot tables. Marketing teams would spend hours downloading data from various platforms, cleaning it, normalising it, and piecing it together into coherent reports. These spreadsheets became the de facto standard for reporting, not because they were the best solution, but because they were the only tool flexible enough to handle the rapidly evolving needs of performance marketing teams.

The weekly marketing review transformed into an elaborate ritual. Mondays became "reporting day"—a frantic rush to update spreadsheets, cross-check numbers, and prepare presentations. Channel managers would work late into the night, ensuring their numbers were accurate and their performance narrative was clear. Tuesdays became a test of endurance: defending performance metrics, explaining cross-channel effects, and justifying budget allocations. (We explore why marketers keep going back to spreadsheets in this post)

Phase 5: When Spreadsheets Stop Working

This was when the first signs of a reporting split began to appear. We found ourselves creating multiple versions of the same reality: daily tracking sheets for optimisation, weekly reports for team reviews, monthly summaries for budget planning, and quarterly decks for executive presentations. Each version required different levels of detail and different ways of looking at the same data. The optimisation sheets tracked granular metrics like keyword performance and audience engagement. The management reports focused on high-level KPIs and trend analysis.

What made this situation particularly challenging was that context was everything. When a campaign underperformed, was it because of increased competition, seasonal effects, or technical issues? The marketer running the campaign would know, but that context often got lost in the reporting process. You might remember that conversion rates dropped because of a two-hour website outage, but three months later, when reviewing quarterly performance, that crucial context would be missing from the spreadsheet.

The spreadsheet era wasn't ideal, but it worked—barely. Marketing teams became spreadsheet wizards out of necessity, building increasingly complex systems to track everything from creative performance to budget pacing. But this was just the beginning. As marketing channels grew more sophisticated and organisations demanded more accountability, even these elaborate spreadsheet systems would prove insufficient.

The Limitations of the Modern Data Stack

The modern data stack emerged as a new paradigm that promised to solve our data chaos. Cloud-based data warehouses like Snowflake and BigQuery would store everything, ETL tools like Fivetran would handle data ingestion, and visualisation platforms like Tableau or Looker would make it all accessible. For performance marketers, this seemed like the answer to our prayers: no more downloading reports from multiple platforms or maintaining complex spreadsheets. Everything would flow automatically into a central repository, ready for analysis. Attribution models could span all channels, creative performance could be analysed across platforms, and we could finally build unified customer journey views. The modern data stack promised to give us what we'd always wanted: a single source of truth for marketing performance.

But then something unexpected happened... This technical division soon created an organisational one.

Marketing teams, needing quick access to granular data and flexible reporting, found themselves constantly requesting dashboard updates and new reports from data teams. Each request required explaining marketing context, defining requirements, and waiting for implementation. Data teams, meanwhile, instead of solving complex data problems or building scalable infrastructure, found themselves caught in an endless cycle of report creation and maintenance. The relationship became increasingly strained: marketing teams saw data teams as bottlenecks to their agility, while data teams viewed marketing as a source of never-ending, ad-hoc requests that prevented them from doing deeper technical work.

This tension wasn't anyone's fault. It was a structural problem. The modern data stack, for all its technical sophistication, had inadvertently created a system where neither team could fully succeed. Marketing couldn't move at the speed their channels required, and data teams couldn't focus on the strategic data problems they were hired to solve.

In Part 2, we'll explore how organisations can bridge this divide by building a crucial operational layer between execution and management—and why this might be the key to restoring performance marketing's lost effectiveness. We'll examine how the right combination of specialised technology and organisational change can help marketing and data teams work together more effectively, and what this means for the future of performance marketing.

This article draws from over two decades of experience in performance marketing, witnessing and participating in every major shift in the industry. One of its authors, Arun (CEO of Clarisights) was one of the first advertisers on "The Facebook" in 2005, when the platform still ran Google AdSense ads.