You Can't Optimize What You Can't See.

Sixty percent of retargeting spend at a major educational publisher was hitting the same users twice. Nobody knew. The data existed across more than 20 vendor portals. Nobody had looked at all of it at once.

The same organization had a senior VP of marketing running A/B and multivariate experiments that were moving user registration rates. Her leadership team could not see those results either. The platform that held them cost per seat, and most of her colleagues had no access.

Two visibility problems. Two payoffs when we fixed them.

TWENTY PORTALS, NO SINGLE VIEW

The publisher ran a portfolio of more than 20 distinct web properties. Each carried its own digital marketing operation: web analytics, paid search, display and banner advertising, retargeting across two separate platforms, affiliate programs, email campaigns, and conversion experiments on the registration flow.

Every activity had a vendor. Every vendor had a portal. Every portal had its own login, its own data format, and its own definition of what success meant. None of them talked to each other.

The VP of marketing oversaw all of it.

Executing that work required a focused team. The people actually running campaigns, managing creative, and coordinating with the web pixeling team needed full Adobe access. That made sense. Adobe Test & Target and Omniture were their instruments.

The problem was the licensing model. Every person who needed to see results cost the same as every person who needed to run campaigns. The tools did not distinguish between a viewer and an operator. So access defaulted to the people who needed it most, and the rest of the organization went without.

The team ran its operations in Excel. Not as a workaround. That was simply where work happened: reviews, decisions, analysis. The stack reflected how the organization operated. And it meant a unified view was one well-built add-in away.

THE VIEW WAS ONE ADD-IN AWAY

Mapping the workflow made the scope concrete. The team maintained campaigns across more than 20 distinct web properties simultaneously. Each property had its own vendor relationships across all seven activity types. Getting a full picture meant logging into each portal, exporting data manually, and reconciling systems that had no awareness of each other.

Roughly 70% of the team's workflows already ran in Excel. Building a new tool meant building inside that environment, not asking anyone to leave it.

“Her experiments were moving registration rates. The problem was nobody outside her team could see it.”

AN ADD-IN, NOT A PLATFORM

Getting close to the work came first. That meant sitting with the team running experiments day to day, understanding what the pixeling team needed to execute a test, and working directly with the ad providers to map what data was actually available via their APIs. The tool design came from that, not from a requirements document.

The team was small by design. One person with web and digital advertising experience who could speak the language of the platforms and the people using them. One application developer. Neither was full-time. The engagement ran three months, June through August, at roughly one full-time equivalent of combined effort.

That combination was sufficient because the problem was specific and the environment was defined.

We built two tools, packaged as a single Excel add-in installable from the ribbon.

The first pulled A/B and multivariate test results from Adobe Test & Target and Omniture directly into Excel. The experiments were specifically focused on the new user registration flow, testing variations across multiple properties to find what converted. Colleagues who previously needed an Adobe seat to view those results no longer did. Only the people actually executing campaigns and coordinating with the pixeling team kept their Adobe access. Everyone else got the information they needed at no additional seat cost.

The second scraped retargeting data from both vendor platforms via their APIs, normalized the output into a common schema, and surfaced it in the same environment. Two platforms, 20-plus individual portals, one interface.

WHAT BECAME VISIBLE

The retargeting consolidation produced the most immediate result. For the first time, data from both platforms sat in the same view. That view revealed something nobody had been able to see before: roughly 60% of retargeting impressions were targeting the same users across both platforms at the same time.

The publisher had two separate vendor relationships, each running its own audience lists, with no mechanism to check for overlap. It was not a vendor problem. Each platform was doing exactly what it was contracted to do. Together, they were doing it twice.

The fix was not technical. It was organizational: work with both retargeting platforms to deduplicate their audience lists. That conversation cannot happen until someone has the proof. The proof required a single view.

The experiment tracker solved the second problem. But the outcome here was not purely operational. The VP had been running registration experiments that moved the numbers. Her leadership team had not been able to see what she was doing or why it was working. Every result she had was locked behind a platform most of her peers had never logged into.

The consolidated view changed that. She could now share campaign performance across the organization without distributing Adobe seats. Leadership could see the lift in registration conversions. She had concrete evidence for budget conversations, planning reviews, and team evaluations.

She had the results. Now she had the receipt.

WHAT THIS MEANS FOR OTHERS

A VP was running experiments that moved registration rates. Her leadership team had no idea. The proof was locked behind a platform most of them had never logged into.

The retargeting duplication was the same problem from a different angle. Two vendors, both doing their jobs, both billing for it. Nobody had seen the overlap because nobody had ever looked at both datasets at once.

Neither required a new platform to fix. They required a single view. The Excel add-in was the right delivery mechanism because 70% of the team's workflows were already there. It landed on their existing desk and worked the day it shipped. Nobody had to change how they worked to use it.

The instinct facing fragmented data is usually to buy something. This engagement ran three months with two part-time people. What unlocked it was a scope decision, not a budget one: one view, built where the team already worked, that made the invisible visible.