The invisible work of turning decisions into tickets, requirements, updates, and alignment

For most of my career, I treated translation as part of being a good product leader.

You listen to the customer and translate what they mean for the product team. You translate an engineering tradeoff for an executive. You turn a planning conversation into requirements, tickets, and a roadmap update. You explain the same decision differently to design, sales, support, and the board.

The best product managers become exceptionally good at this. They learn each audience’s language, anticipate what context is missing, and carry information between parts of the company that otherwise struggle to understand one another.

We often describe that as being the glue. While building Earmark, we started to see the less flattering version: the organization was using people as middleware.

One of the most expensive sentences in product work may be: “I’ll turn this into something everyone can use.”

The sentence sounds responsible because it is. But hidden inside it is another job: reconstruct the conversation, identify what mattered, decide what each audience needs, rewrite it in the appropriate format, enter it into several systems, and answer the clarification questions that follow.

We came to think of this as the translation tax - the recurring labor required to convert what an organization already knows into a form that allows someone else to act.

Translation is everywhere, but almost nowhere on the calendar

The translation tax rarely appears as a scheduled activity. A calendar records the product review but not the decision log written afterward. It records the engineering discussion but not the tickets created from it. It records the customer call but not the work required to turn what was learned into evidence the broader team can evaluate.

Because the work is distributed across dozens of small actions, it is easy to underestimate. Reply to the thread. Update the document. Copy the decision into Jira. Rewrite the technical explanation for leadership. Find the previous conversation. Ask whether the requirement changed. None of these actions looks especially costly by itself.

Together, they become a large share of the workday.

More than a decade ago, the McKinsey Global Institute estimated that interaction workers spent 28% of the workweek managing email and nearly 20% searching for internal information or finding colleagues who could help. The specific tools have changed since that research was published, but the underlying activity is familiar: locate the context, understand it, and move it somewhere useful.

Newer research suggests the burden has not disappeared. In a 2024 survey of more than 10,000 desk workers, Slack’s Workforce Lab found that people reported spending 41% of their time on tasks they considered low-value, repetitive, or disconnected from their core job.

Some coordination will always be necessary. The problem is how much of it exists only because our systems cannot carry meaning across the boundaries between conversations, people, and tools.

The translation tax is not the cost of creating new knowledge. It is the cost of making the same knowledge usable again and again.

Moving information is not the same as carrying meaning

Modern companies are very good at moving information. A transcript can be stored instantly. A document can be shared with the entire organization. A Jira ticket can link to a Slack thread, which links to a customer recording, which links to a roadmap item.

But links do not perform the translation.

Someone still has to decide which part of the customer conversation changes the requirement. Someone has to explain why engineering rejected one approach and what that means for the launch. Someone has to turn a nuanced discussion into a ticket concise enough to use without stripping away the constraints that made the decision sensible.

This is semantic work, not clerical copying. The same conversation may need to become a technical requirement, an executive summary, a customer follow-up, and a set of tasks. Each artifact expresses the same underlying context for a different purpose.

The work also crosses applications constantly. Research described in Harvard Business Review followed 137 people across 20 teams at three Fortune 500 companies. Workers toggled between applications and websites roughly 1,200 times per day, losing just under four hours each week to reorienting after those switches - about 9% of their annual work time.

That measurement captures the visible switching cost. It does not fully capture the cognitive work of changing audiences and representations: from listening like a product manager to writing like a business analyst, then communicating like an executive, then organizing work like a project manager.

The costly switch is not only from Zoom to Jira. It is from what people meant to what every other person and system needs to understand.

The tax compounds as context moves

Translation does not only consume time. Each handoff creates an opportunity for meaning to change.

A customer describes a problem. A product manager summarizes it. A team turns the summary into requirements. Engineering converts the requirements into an implementation plan. Leadership receives a status update several steps removed from the original conversation.

At every stage, someone makes reasonable decisions about what to include, simplify, or omit. But by the time the work reaches implementation, the team may remember what it is building without preserving why the original problem mattered.

This is why translation quality affects more than efficiency. Project Management Institute research found that organizations rated as highly effective communicators were much more likely to meet original project goals, deliver on time, and stay within budget than minimally effective communicators. PMI has also reported that 47% of unsuccessful projects failed to meet their goals because of inaccurate requirements management.

Those findings are often interpreted as a need for better documentation and communication discipline. That is part of the answer. But discipline alone does not solve the structural issue: the more often humans must manually reinterpret context, the more opportunities there are for drift.

The safest translation is the one produced while the context is still present, with the people who understand it available to review it.

Earmark’s discovery was about completion, not capture

We did not begin Earmark with the phrase “translation tax.” We arrived there by watching what people tried to do with the conversation after it happened.

Across more than 100 discovery conversations with product managers and product leaders, people rarely described their ideal outcome as a better transcript. They wanted the work on the other side of the transcript: the decision log, the stakeholder update, the requirements, the Jira tickets, the follow-ups, and the preserved rationale they would otherwise have to recreate later.

Our own product evolution followed that discovery. Transcription and summarization were useful, but they did not remove the translation layer. They created better raw material for the person still responsible for turning the conversation into work.

That changed our bar. The question was no longer, “Did we capture what happened?” It became, “What work no longer has to happen because the conversation was captured?”

A transcript saves memory. Finished work saves time.

AI can make the tax cheaper - or make it disappear

Generative AI can already accelerate individual acts of translation. Paste a transcript into a model and ask for a status update. Turn meeting notes into tickets. Rewrite a technical explanation for an executive audience.

That workflow is useful, but it often leaves a person responsible for assembling the context, writing the prompt, checking the output, choosing the format, and moving the result into the appropriate system. The translation becomes faster without becoming automatic.

It can also increase output without reducing coordination. If everyone can generate more documents, summaries, messages, and tickets, the organization may end up with more material to reconcile and more questions about which version is authoritative.

The larger opportunity is not a faster translator sitting beside the workflow. It is a work layer that understands the conversation, preserves its context, produces the appropriate artifacts, and allows people to review them before they move into systems of record.

Humans should still decide, refine, and approve. But they should not repeatedly reconstruct the same source material for every audience and destination.

Product teams should spend their judgment, not their time

Great product work will always require translation in the deeper sense. Customers, engineers, designers, and executives see the world differently. Helping them understand one another requires judgment, empathy, and taste.

What should disappear is the mechanical burden surrounding that judgment: recovering the conversation, reformatting it, copying it between systems, and reproducing context that the organization already had.

That is what compels us to build Earmark. Product managers should be translators of meaning when judgment matters—not full-time converters of meetings into documents. The value is in deciding what matters, not in manually manufacturing every artifact required to carry that decision forward.

The tools of knowledge work have spent years making it easier to communicate. The next generation must make it unnecessary to translate every communication into work by hand.

The translation tax disappears when conversation and execution become the same workflow.

Sources

• McKinsey Global Institute, “The social economy: Unlocking value and productivity through social technologies”, July 2012.

• Harvard Business Review, “How Much Time and Energy Do We Waste Toggling Between Applications?”, August 29, 2022.

• Slack Workforce Lab, “New Slack research shows accelerating AI use and quantifies the ‘work of work’”, February 27, 2024.

• Project Management Institute, “Communication”.

• Project Management Institute, “Requirements Management: Core Competency for Project and Program Success”, August 2014.