Coordination Consumes 35 to 60% of Your Team's Time. Change Futurist | September 21, 2026
Count the handoffs in your highest-volume marketing workflow. Most teams have never measured that number, and it caps what they get back from every model they license.
Four research groups published inside five days, working from different samples and different agendas, and all four describe the same failure point. Ali Sankur, Bernhard Mühlreiter, and Steffen Fuchs at McKinsey put a price on the work that happens between process steps. Anurag Raj at Gartner reports that cultural resistance outranks funding as the reason governance programs fail. Molly Innes reported research finding no chief marketing officer on the main board of any FTSE 100 company. MarketingProfs editors compiled a week of launches aimed at buyers who arrive through an agent.
You lose the value at the interfaces, and there is a person on each side of every one of them.
Last week I wrote about assigning authority to each step in a workflow. That work matters, and it handles only part of the problem. A workflow with clean decision rights on every step and nine unmanaged handoffs between them still runs on the queue.
Count the handoffs before you count the use cases
Sankur, Mühlreiter, and Fuchs published Cutting the 'coordination tax': How agentic AI can reshape workflows on September 18. Their central measurement belongs in your next operating review: coordination, which they define as aligning, verifying, reconciling, waiting, and handing work from one step to the next, consumes 35 to 60 percent of total work time in knowledge-intensive organizations.
They break the cost into three layers. Visible costs cover the people and structures dedicated to coordination, running 3 to 5 percent of revenue in the manufacturing organizations they studied. Latency costs cover the capital and capacity trapped while work waits. Opportunity costs from slow decisions run largest of the three.
Then the history. The authors walk through five improvement eras, from Lean and Six Sigma through business process reengineering, ERP, process management, and robotic process automation. Each one optimized inside process steps. Each one left the interfaces between steps untouched.
The diagnostic number from their Fortune 500 manufacturer case is the one to steal. Actual processing time ran 12 to 24 hours. Interface latency ran 9 to 18 days. That gap runs nine to thirty-six times, at an estimated coordination tax of $140 million to $240 million a year. After the client redesigned the interfaces, the planning cycle went from 30 days to 3 and 80 percent of decisions became touchless.
Their investment split reads 70 percent organizational change, 20 percent technology and data, 10 percent model development. Marketing leaders who have been inverting that ratio for three years now have a benchmark to argue with.
It reminds me of what Jessica Vose, Head of Marketing at Kana, said when I interviewed her on The Agile Brand podcast: "The marketing cloud and the CDP both sold coordination and both mostly delivered consolidation; for example, a single vendor's login screen instead of six, a single database instead of six. That's progress, but it isn't the same job as coordination."
I have called this structural lag: the distance between knowing and doing. A model scores the right next action in milliseconds while your organization spends four days approving it. Vendors sell business orchestration and automation technologies to close that distance, and buying one without redesigning the interfaces gives you a faster way to wait.
My take: Sankur and his coauthors supply the measurement marketing operations leaders have lacked, because a handoff count and an interface latency figure convert an abstract complaint about silos into two numbers a CFO can act on.
Your governance problem is a trust problem
Raj published a prediction on September 21 from the Gartner Data and Analytics Summit in Mumbai: 60 percent of organizations that ignore data governance culture challenges will fail to govern AI successfully by 2027. Across 223 data and analytics leaders surveyed in March 2026, cultural resistance outranked funding constraints as the reason governance programs fail, 60 percent against 40 percent.
Raj states the operative principle directly: "AI-ready data also requires AI-ready stakeholders who understand the value of trusted data."
Read that against the coordination tax and the mechanism becomes concrete. A handoff carries data across an organizational boundary. The person receiving it either trusts what arrives or re-verifies it, and the re-verification is the latency. A team re-derives numbers another team already produced because someone got burned once and told the story at an offsite, and no amount of pipeline engineering fixes that.
Gartner analysts put the stakes plainly in the same event's Day 1 summary. In 2025, the odds of an AI initiative achieving a return on investment ran one in five.
It reminds me of what Prachi Gore, Chief Marketing Officer at Asana, said when I interviewed her on The Agile Brand podcast: "I think the biggest change in mindset is we have to think agents-first in everything… The second mindset shift is, at every level in the company, you've all become managers now."
Managers own the quality of what they hand to the next person. Raj is describing that same obligation from the data side.
My take: Raj puts a survey figure behind something operators learn the hard way, and on a 60 against 40 split, your next governance dollar belongs with the analyst who refuses to trust the shared definition rather than with another cataloging tool.
Check where your marketing operations leader reports
Innes reported on September 18 that research from The Marketing Directors found no marketing directors or CMOs on the main board of any FTSE 100 company, down from 14 board-level marketing officers in 2007. Guy Tomlinson, chief marketer at The Marketing Directors, credits marketing's rising executive importance to "the ability to understand customer intent, use data and technology intelligently." Marketing expertise now sits in executive committees and operating boards below the main board: 45 percent of main boards hold marketing expertise beyond the chief executive, and 30 percent have senior marketing leaders one level down.
Put that next to what employers are paying for. Innes reported on September 17 that Michael Page's 2027 Salary Guide names AI adoption as the most in-demand marketing skill for 2027, followed by marketing automation and customer journey management, with 77 percent of marketers using AI tools daily. That list describes a marketing operations competency stack, priced by the people doing the hiring.
So the function acquiring the capability to redesign commercial interfaces holds no seat in the room where the board allocates capital. A director can map handoffs inside marketing. Clearing the interface between marketing, service, and product takes someone who can direct three budgets, and that person sits on an operating board reporting upward.
This is why I keep intelligence, operations, and experience as horizontal capabilities in the framework I am building on top of my upcoming book, Coherence at Velocity. Horizontal means the capability runs across every team, which means the owner needs standing across every team. Give a director a cross-functional mandate and a functional budget and you have given them a calendar full of meetings.
My take: Innes documents a seventeen-year decline in formal marketing authority while customer-intent data becomes a primary input to enterprise AI strategy, so set the reporting line of your marketing operations leader with the same care you give your platform roadmap.
Design for the buyer who never reaches your site
MarketingProfs editors compiled roughly thirty developments from the prior week in their AI Update for September 18. Three of them change where you hand a buyer off.
OpenAI began testing Sponsored Agent ads that route a shopper into a branded conversation inside ChatGPT, with Wayfair participating. Brands including Time and Wayfair are testing formats aimed at autonomous systems rather than people. Visa, Mastercard, and Ant are building a know-your-agent identity framework for payments, which means the agent transacting on a customer's behalf will carry credentials your systems have to recognize.
The demand-side figure that should move your Q4 plan comes from Shopify: shoppers who begin product searches through large language models are 2.5 times more likely to land directly on a product page, contributing to roughly an 80 percent increase in conversion. A separate estimate in the same roundup puts automated and AI-directed campaigns at 12 percent of US ad spending this year, up from 2 percent in 2023, with a projection of 27 percent by 2030.
Each of those items creates an interface, and nobody has assigned the handoff yet. Your product data crosses into an engine you cannot observe. Customer journey orchestration built for human sessions routes an agent the way it routes a browsing shopper, so your pacing logic and your abandonment triggers fire against something that never reads them.
Assign those handoffs now, while the volume stays small enough that a mistake costs you one quarter.
My take: The AI Update reads as an infrastructure bulletin this week, since agent identity standards and agent-directed ad formats both move the point of customer contact outside the systems marketing teams currently instrument.
Price the handoffs in a workflow you recognize
A $1.2B consumer insurance carrier runs renewal retention with a team of forty across marketing, operations, and service. The workflow that produces a renewal offer has nine handoffs in it: risk scoring to pricing, pricing to offer construction, offer to legal review, legal to creative, creative to brand review, brand to channel build, build to data ops for audience assembly, data ops to deployment, deployment to service for inbound handling.
Actual work time across those nine steps runs about 38 hours. Calendar time from risk score to customer inbox runs 26 days. The team has spent eighteen months adding models to four of the nine steps, and calendar time has moved from 29 days to 26.
Now price the interfaces. Three of the nine handoffs carry the cost. Legal review waits an average of six days because requests arrive as email attachments with no standard intake. Audience assembly waits five days because data ops re-derives the eligibility logic pricing already applied, since the two teams use different definitions of an active policy. Brand review waits four days because reviewers receive assets one at a time and batch them by habit.
Each repair is organizational. Legal gets a structured intake form and a two-day service level for anything inside a pre-approved claim library. Pricing and data ops agree on one definition of an active policy, write it down, and data ops stops re-deriving. Brand review moves to a standing daily slot with a rule that silence past the slot counts as approval on templated derivatives.
Calendar time goes from 26 days to 11. The four models the team already licensed now run inside a cycle short enough that a risk score stays current when the offer reaches the customer. The carrier can send retention offers monthly instead of quarterly, which moves the addressable retention math further than any model upgrade on the roadmap.
Forty people and four models, unchanged. The repair cost three meetings and one written definition.
Take these four steps into Q4 planning
Map and time-stamp the handoffs in your three highest-volume workflows. Record actual work time and calendar time separately for each step. The ratio between them is your coordination tax, and McKinsey's nine-to-thirty-six-times range gives you something to compare against.
Spend the first governance dollar on the re-verification habit. Find every place one team re-derives a number another team already produced, and fix the definition and the trust deficit behind it. On Raj's 60 against 40 split, the cultural work returns more than the tooling.
Name who owns the interfaces between marketing, service, and product. A functional director cannot clear a cross-functional queue. Put the mandate where the three budgets meet, and write down what that person can decide without escalating.
Instrument the agent-mediated handoff before Q1. Separate agent traffic from human sessions in your analytics, confirm your product data reads cleanly to an engine, and decide now who owns accuracy when an agent transacts on a customer's behalf.
I have written more than a couple dozen books, and I can say that drafting a single chapter can take days. Producing a manuscript takes the better part of a year, and almost all of that year is queue: a chapter sits with a developmental editor, comes back, sits with a copy editor, sits with a permissions check. On one book I cut the calendar by a third. Nothing about my drafting speed changed. We put every reviewer into one shared document, gave each a standing weekly slot, and agreed that silence past Friday counted as approval.
Count your handoffs. Then time them, price the worst three, and fix those before you license anything else.
Featured insights:
Cutting the 'coordination tax': How agentic AI can reshape workflows. Ali Sankur, Bernhard Mühlreiter, and Steffen Fuchs, McKinsey, September 18, 2026
Gartner Predicts 60% of Organizations That Ignore Data Governance Culture Challenges Will Fail to Govern AI Successfully by 2027. Anurag Raj, Gartner, September 21, 2026
No CMOs on top board of any FTSE 100 company, study finds. Molly Innes, Marketing Week, September 18, 2026
AI Update, September 18, 2026: AI News and Views From the Past Week. MarketingProfs, September 18, 2026
Also referenced: Senior salaries up 40% outside capital as London wages flatline, Molly Innes, Marketing Week, September 17, 2026, and Gartner Data & Analytics Summit 2026 India: Day 1 Highlights, September 21, 2026.