Fixing a Complex Client Delivery Process With a Custom Operating Platform

Overview

As businesses grow, their most important processes often become their messiest.

What starts as a manageable workflow becomes a patchwork of spreadsheets, email threads, project boards, manual checks and individual habits. The work still gets done, but only because experienced people know where everything lives, what needs checking, and who to chase next.

That was the situation Link had reached with one of its most important client delivery processes.

The process was essential, touched multiple teams, and had grown in complexity over time. But the systems supporting it had not kept up. Link had explored off-the-shelf SaaS tools, but none matched the way the work actually happened. Some were too generic. Some were too rigid. Some were too expensive to roll out properly across a large team.

So we helped Link do something different: build a custom operating platform around the process itself.

Who are Link?

Link is a global marketing procurement agency working across 60+ markets for some of the world's biggest brands.

They manage complex client programmes involving multiple teams, suppliers, markets and stakeholders. Delivery quality matters, visibility matters, and small operational issues can quickly become client-facing problems.

The Challenge

Link did not have a simple software problem. They had a process problem.

The work was important, but the way it was managed had grown organically over time.

The process depended too heavily on individual knowledge. Different team members had developed their own ways of working. Some things lived in spreadsheets. Some lived in project tools. Some lived in inboxes. Some lived in people's heads.

That created fragility. When someone experienced was involved, the process worked. But when people were unavailable, moved role, or left the business, too much knowledge went with them.

There was no single source of truth. Different teams were often working from different versions of the same information.

That creates the usual problems: duplicated effort, outdated data, inconsistent status updates, avoidable mistakes and too much time spent checking what is actually true. When a process is small, this is annoying. When it is business-critical and touches major clients, it becomes a serious operational risk.

Handoffs were too manual. The process moved across multiple teams and stages. Each handoff relied on people knowing what had happened before, what needed doing next, and what information mattered.

That meant the team spent too much time coordinating the work around the process, rather than simply doing the work.

Scaling the process meant adding people. Because so much of the workflow was manual, growth created more admin.

More clients, more campaigns, more checks and more edge cases meant more people were needed to hold the process together. That is not a scalable operating model.

Off-the-shelf tools did not fit. Link had looked at generic SaaS and project management platforms. The problem was not that these tools were bad; it was that they were not designed for this process.

To make them work, Link would either have had to simplify the process around the tool, accept awkward compromises, or pay for a large number of seats across the business. None of those options solved the real problem.

The Insight

The key insight was simple: this process was too important to keep forcing into tools that were not built for it.

Link did not need another spreadsheet, another project board, or another generic SaaS subscription. They needed a system that reflected how the work actually moved through the business.

That meant starting with the process, not the software.

What We Did

1. Mapped the process end to end

We worked with Link to understand the full workflow in detail: who is involved, what happens at each stage, what information is needed, where decisions are made, where approvals happen, what handoffs exist between teams, what data changes over time, and where mistakes and delays typically appear.

This was not about documenting an idealised process. It was about understanding how the work actually happens. That mapping gave us the blueprint for the platform.

2. Designed the right operating model

Once the process was clear, we translated it into a software model.

That meant defining the core objects, statuses, responsibilities, permissions, gates, handoffs and views the team needed.

The goal was not to replicate the old spreadsheet process in a web app. The goal was to design a better operating system for the work.

3. Built a custom internal platform

We then built a dedicated platform around Link's workflow.

The platform gives teams one shared place to manage the process, track progress, update status, handle approvals, assign ownership and see what needs attention.

Instead of relying on scattered files and individual memory, the process now lives inside a structured system.

4. Iterated with real users

Because the process was complex, some requirements only became clear once people started using the system.

We worked closely with Link's team, took feedback from real users, and kept improving the platform around the way the work actually needed to happen.

That included better views, clearer handoffs, stronger controls, improved data structure, more useful notifications and new workflow steps as the process matured.

The Result

Link now has a custom operating platform for a process that was previously held together by spreadsheets, manual coordination and individual expertise.

One source of truth. Teams can work from the same live information, rather than checking multiple files or relying on outdated updates. That reduces confusion, avoids duplicated work and gives everyone a clearer view of what is happening.

Easier handovers across teams. The process is no longer dependent on one person knowing what happened last. Status, ownership, history and next actions are visible in the system, making it much easier for work to move between teams.

Less reliance on institutional knowledge. Experienced people still matter, but the process no longer depends on everything being stored in their heads. The platform captures the workflow, the data and the decision points, making the business more resilient as people move roles, take leave or leave the company.

Better client visibility. Because the process is more structured, Link can give clients clearer updates and better visibility. That improves confidence and reduces the friction that comes from slow responses, unclear statuses or inconsistent information.

A more scalable cost model. Instead of relying on expensive per-seat SaaS tools across a large team, Link now owns a system built around its process.

That matters when software needs to be used by many people across the business. Per-user pricing can become expensive quickly, especially when the tool still does not properly fit the workflow.

A foundation for AI and automation. The custom platform also creates the foundation for future automation.

Once the workflow, data and handoffs are structured, AI can be added in useful ways: generating summaries, flagging risks, prompting next actions, automating follow-ups, surfacing exceptions and reducing manual admin.

The important point is that AI is being added to a proper operational foundation, not layered on top of a messy process.

Why This Matters

This is the kind of problem many growing businesses have.

They do not always need a full enterprise software transformation. They do not always need another SaaS platform. And they definitely do not need to keep forcing complex work into spreadsheets forever.

They need someone to understand the process, design the right system, and build practical software around the way the business actually works.

That is what we did for Link.

Key Outcomes

• Replaced scattered spreadsheet-led tracking with a dedicated internal platform
• Created one source of truth for a complex client delivery process
• Improved handovers across teams
• Reduced reliance on individual institutional knowledge
• Improved client visibility and confidence
• Avoided forcing the workflow into expensive generic SaaS tools
• Created a foundation for further AI and automation