A rep leaves the depot at 07:40 with a perfect day in her pocket. Six stops, 143 km, an optimizer-approved sequence, 4 hours and 20 minutes of face time. It's the kind of plan that looks great in a QBR deck.
At 10:47 it's dead.
Stop two — a mid-sized distributor, her third-largest account in the territory — has the operations manager stuck in an unscheduled quality meeting. She waits 25 minutes in reception, gets 12 minutes instead of 40, and walks out at 11:31 with a half-finished conversation and a day that no longer exists. Stop three starts lunch. Stop four gets compressed. Stop six gets a text message: "Sorry, can we push to next week?"
Nothing about the routing was wrong. The sequence was optimal. The distance was minimal. The plan was simply brittle — and brittleness, not distance, is what quietly destroys field-sales capacity.
We optimize the plan. Nobody engineers the recovery.
Ask a sales ops team how they evaluate a tour plan and you'll hear about kilometres, drive time, stops per day, maybe fuel. All of those measure the plan as designed. None of them measure what happens after reality shows up.
And reality shows up daily:
- The contact is on site but unavailable
- A 30-minute check-in turns into a 90-minute opportunity you'd be insane to cut short
- An accident closes the A5 and adds 40 minutes
- An existing customer calls with an escalation that needs a same-day visit
- The 16:00 appointment cancels while you're still at the 11:00
The right question isn't "what's the shortest route through these six stops?" It's "when this plan breaks — and it will, maybe twice a week — how much of the day do we get back?"
That's a design problem, not a math problem. Here's how to solve it.
1. Build days around anchors, not sequences
A fully optimized six-stop chain is a series of dependencies. Break link two and links three through six all move.
Instead, structure the day as two or three anchors plus flexible fill:
- Anchor A: a confirmed appointment before 10:00
- Anchor B: a confirmed appointment between 14:00 and 15:30
- Fill: three to four accounts that sit within ~10 minutes of the corridor between them, visitable without an appointment
Anchors hold the geography of the day in place. Fill absorbs the variance. When stop two blows up, the rep isn't rebuilding a day — she's swapping one flexible stop for another inside a corridor she's already driving.
This also fixes a subtle failure: reps who chase perfect sequences tend to book appointments in whatever slots customers offer, which scatters the anchors and makes every day fragile from the start. Booking discipline — "I need Tuesday morning in the Rhein-Neckar corridor, not just Tuesday" — is a routing decision disguised as a calendar decision.
2. Density is insurance, not just fuel savings
Everyone justifies tight territories with drive time. The bigger payoff is optionality.
If a rep's day corridor has 30 relevant accounts within 15 minutes of it, a collapsed stop costs 20 minutes. If the corridor has four, it costs the afternoon.
So when you redraw territories, don't only balance account count and revenue potential. Add a third test: can a rep recover from a failure anywhere along this route without driving more than 15 minutes? A territory that passes that test tolerates chaos. One that doesn't will look efficient on paper and bleed hours in practice.
Practical version: every planned day should ship with a standby list of five accounts — nearby, low-ceremony, genuinely worth a visit. Not filler for filler's sake. Accounts that are overdue on cadence, or where a drop-in has a real chance of moving something. If your CRM can't generate that list in one click for a given corridor, your reps will improvise it badly, or go home early.
3. Put buffer where the variance actually lives
The lazy fix is padding every stop by 15 minutes. That destroys capacity — six stops, 90 minutes gone, and you've solved nothing because the variance isn't evenly distributed.
Buffer by stop type, based on your own history:
- First discovery meeting with a new prospect: high variance, both directions
- Quarterly review with a strategic account: high overrun risk, low no-show risk
- Routine restock or check-in with a stable account: low variance, tight window
- Cold drop-in: near-zero overrun risk, high no-show risk — perfect as fill, terrible as an anchor
The data to do this already exists in your check-in timestamps. Actual dwell time per account, per visit type, per rep. Most teams never look at it. Planners like SalesFleet use exactly this — learned dwell times rather than a flat default — because a 30-minute assumption applied to an account that historically eats 55 minutes guarantees a broken afternoon, every time.
If you want one illustrative sense of the stakes: a rep who loses 45 minutes twice a week to plan collapse loses roughly the equivalent of a full selling day per month. That's not a benchmark, just arithmetic on plausible numbers — run it with your own.
4. Give reps a re-plan trigger, not a re-plan tool
Handing a rep a re-optimize button is useless if nobody has told them when to press it. Left alone, most reps improvise: they drive to the next stop on the list because it's the next stop on the list, and absorb the damage silently.
Set an explicit decision rule. Something as blunt as:
If you're more than 25 minutes behind by the third stop, stop driving and re-plan.
Then define what gets cut, in advance:
- Drop the stop with the weakest reason-to-visit this week, not the furthest one
- Never drop an account that's already been rescheduled once — that's how relationships quietly die
- Prefer dropping a stop you'll pass again within 10 days over one you won't
Reps make better calls in 30 seconds at a rest stop than in 3 seconds at a junction. The rule matters more than the software.
5. Make the last stop cheap to lose
Sequence so the most droppable visit sits last and closest to the rep's home. Two benefits: the tail of the day is expendable without geographic penalty, and when the day does hold together, the final leg home is short instead of a 70-minute unpaid drive at 18:15.
Reps notice this one immediately. It's also the easiest to get wrong, because pure distance optimization will happily end the day 90 km from where someone lives.
The four numbers that tell you if your plans are real
Stop reporting only stops-per-day and kilometres. Add:
- Plan survival rate — share of days where every planned stop was completed as planned. If it's under 60%, your plans are fiction and everyone knows it.
- Recovery time — average minutes between a stop collapsing and the next productive activity starting.
- Drop position — which slot in the day gets sacrificed most. If it's consistently slot two or three, your anchors are misplaced.
- Dwell variance by visit type — the raw material for buffer decisions.
One warning: don't turn plan survival into a compliance metric. The moment reps are graded on adherence, they sandbag — four stops planned, four completed, 100% adherence, half a day of capacity gone. Measure the plan's quality, not the rep's obedience.
Try this on Monday
Pull the last 10 days where a rep finished with unplanned stops or an early exit. For each one, write a single line: what broke, at what time, and what the rep did next.
You'll almost certainly find one dominant cause — a specific account that always runs long, a booking habit that scatters anchors, a territory with no recoverable density in its north-east corner.
Fix that one thing. It'll return more selling hours than another round of route optimization ever will.