The Operational Last Mile: The Stretch No System Owns

Every system in a modern operation was bought for a good reason, and most of them do their job. The CRM holds the pipeline. The ERP closes the books. The HR platform runs payroll. Each one is reliable inside its own four walls.

What no system owns is the space between them: where an order becomes an invoice, where a hire becomes a provisioned employee, where a case crosses three teams before it closes. That space is where the operation actually runs, and it runs on spreadsheets, email threads, and someone re-keying data from one screen into another.

Each system works. The operation between them is held together by people. That space has a name: the operational last mile, and it is the least deliberately designed part of most technology estates.

What the Last Mile Actually Is

The term comes from logistics. A parcel crosses the country efficiently on trunk routes, then meets the final stretch to the doorstep: short, unpredictable, and human. That last stretch absorbs a disproportionate share of the total cost and most of the failures.

Operations has the same shape. Your systems of record are the long-haul network: they store data reliably within their own four walls. The last mile is everything that happens between them: the handoffs, the exceptions, the re-keying, the chasing that carry work across system and team boundaries. The distance is trivial on a diagram and punishing in practice, because no system owns it.

Before going further, it helps to see where your last mile lives. It's rarely everywhere at once; it's usually concentrated in one or two flows that quietly run on manual effort.

Locate your last mile

Where does the work actually happen?

Pick the flow that runs most on spreadsheets, email, and re-keying. That's where your last mile lives.

Order-to-cash

Quote → order → fulfil → invoice → cash

Systems you've already bought

CRMERP / financeFulfilmentBilling

Where the last mile hides

Orders re-keyed from email or CRM into the ERP, invoices reconciled by hand against what actually shipped, and cash chased from a spreadsheet.

The tell

No one can tell you the true status of an order without opening three systems and asking two people.

What a connected layer changes

One operational record follows the order end to end, so status is live, invoicing matches fulfilment automatically, and collections work from the same truth as sales.

Wherever you landed, the shape is the same. The systems on either side do their job; what happens between them is improvised.

Why the Last Mile Stays Unpaved

The last mile persists for a structural reason, not a lack of effort: systems of record are built to store data, not to move work across each other. Each one optimises its own four walls and treats everything beyond them as someone else's problem. The seams between systems are unowned by design.

The scale of that fragmentation is now extreme. The MuleSoft 2025 Connectivity Benchmark puts the average organisation at 897 applications, yet only 2% have successfully integrated more than half of them. The rest are islands. Unsurprisingly, 90% of IT leaders say data silos are creating business challenges, and 95% report struggling to integrate data across systems.

Every unintegrated boundary is a stretch of last mile. And because each boundary looks small on its own, no one ever funds paving it — they just keep paying to walk it by hand.

The Cost Hides in Plain Sight

The reason the last mile escapes scrutiny is that it never appears as a line item. It's absorbed into salaries and lost time, which makes it feel free. It isn't.

Some of it is the toggle between systems. Harvard Business Review's study of workers across three Fortune 500 companies found the average person toggles between applications nearly 1,200 times a day, losing just under four hours a week, about 9% of their time at work, simply reorienting after each switch.

Some of it is hunting for information that already exists. The McKinsey Global Institute found the average knowledge worker spends nearly 20% of the workweek looking for internal information or tracking down colleagues, and 28% managing email — before any real work begins.

And some of it is the cost of systems that quietly disagree. Gartner estimates poor data quality costs organisations at least $12.9 million a year on average — much of it the direct product of the same record living differently in every system, reconciled by hand before anyone can act.

None of these are technology bills. They're the running cost of an unpaved last mile — and they recur every single day.

Why the Usual Fixes Miss It

Faced with the symptoms, most organisations reach for one of three fixes — and each misses the last mile in its own way.

More software adds another system of record, and with it another boundary to cross. It lengthens the last mile rather than paving it. (This is how you reach 897 applications — the subject of Platform vs. Point Solutions.)
More integration connects the APIs so data can move, but it doesn't carry how work moves — the ownership, the handoffs, the exceptions. Connected data with no workflow around it is a faster island. (We quantify this in The Integration Tax.)
More people — the default when a process breaks — just staffs the manual last mile. It scales the cost linearly and hard-codes the fragmentation into headcount.

Each treats the last mile as a technology gap or a labour gap. It's neither. It's an operational-architecture gap — the missing layer that sits across the systems and owns the movement of work between them.

Paving It Is a Progression, Not a Project

Here's the trap: because the last mile feels like a gap, leaders try to close it in one leap — a big integration programme, a new platform, a transformation. But you don't pave the last mile with a project. You advance along a spectrum of operational-integration maturity, one flow at a time.

Find where you sit honestly before you decide where to invest.

Operational-integration maturity

How paved is your last mile?

Drop a marker on the stage that matches your operation today — not where you'd like to be.

FragmentedConnected

Stage 1

Manual last mile

The systems exist; the operation runs on spreadsheets

What it looks like

Work moves between systems by hand. People re-key data, chase status over email, and keep the process alive in personal spreadsheets and checklists.

The tell

The real operating system of the business is Excel and someone's inbox.

The cost you're still paying

The full integration tax — re-keying, reconciliation, latency, and key-person risk — paid every day and mostly invisible.

The next move

Map where work actually breaks down before buying anything. Start with the one flow that hurts most and make its handoffs visible.

Most organisations are further left than they'd like to admit — and the jump that matters is rarely the whole distance. It's the next stage: making one flow's handoffs explicit, or giving one process a shared record. Progress compounds from there.

What "Paving" Actually Means

Advancing along that spectrum doesn't mean ripping out systems. It means building the layer that's been missing — an operational layer that sits across your existing tools rather than adding another one beside them. Concretely, it does four things no system of record does on its own:

A shared operational record — one version of the customer, order, or employee, so reconciliation stops being a job.
Workflow that spans systems — explicit handoffs, clear ownership, and visible status as work moves end to end.
Exception handling by design — the hard cases routed to the right owner instead of parked in an inbox.
A live operating picture — the single view assembled once, in real time, instead of rebuilt by hand each period.

The systems don't change. The operation does. (Whether to build that layer or buy it is its own decision — we work through it in The Build vs. Buy Decision.)

The Path Forward

The operational last mile is where strategy meets reality: it's the difference between the operation you designed — systems talking, work flowing — and the one you actually run, held together by spreadsheets and goodwill. It stays invisible precisely because it hides inside salaries and month-end scrambles rather than on an invoice.

The organisations that keep ignoring it will buy more software, connect more APIs, and hire more people — and wonder why the operation still feels manual. The ones that pave it will start where the evidence points, close one flow's last mile at a time, and build the operational layer that carries work across their systems instead of leaving it to people.

The starting point is always the same: find where your last mile actually lives — then pave the stretch that's costing you most.

Want to find and close your operational last mile? Start the conversation — we'll help you map where work breaks down and show you where a connected layer would pay back first.

Sources

Salesforce / MuleSoft. 2025 Connectivity Benchmark Report. 2025. MuleSoft blog summary — average of 897 applications per organisation; only 2% have integrated more than half of their applications. Salesforce newsroom — 90% of IT leaders say data silos are creating business challenges; 95% report struggling to integrate data across systems.
Harvard Business Review. How Much Time and Energy Do We Waste Toggling Between Applications? August 2022. hbr.org — across 20 teams (137 users) at three Fortune 500 companies, workers toggled roughly 1,200 times per day and lost just under four hours per week (≈5 working weeks, or 9% of annual time) reorienting after switches.
McKinsey Global Institute. The Social Economy: Unlocking Value and Productivity Through Social Technologies. July 2012. mckinsey.com — the average interaction worker spends an estimated 28% of the workweek managing email and nearly 20% looking for internal information or tracking down colleagues.
Gartner. Data Quality: Why It Matters and How to Achieve It. Gartner research (2020). gartner.com — poor data quality costs organisations at least $12.9 million a year on average.