Algeria AgencyIndustries
The round: building a driver’s day, and what happens when an address fails
You are paid per parcel and you spend per attempt. A driver’s whole day is decided in that gap.
The article beside this one is about the network: the lead-time promise, the address, tracking, coverage, remittance. It describes what happens between a company and its customers.
This article is about one day. One person, one vehicle, one load, one city, and between forty and a hundred doors — a number of which will not open.
The central problem is a mismatch of units, and it explains almost everything else: you are paid per parcel and you spend per attempt. A parcel delivered on the third pass earns once and costs three times.
We describe the order of the round, what happens at a closed door, the rule for the second attempt, and the five minutes at the end of the day. We give no success rate: section 10 says why.
A round is not a list of addresses
A list of addresses is what the system prints. A round is what a person will do with that list between eight and five, with a vehicle, in a city they know better than the system does.
The difference is three pieces of information the list does not contain: how long each stop takes, the probability that the door opens at that hour, and what the journey between two points that look close on a map actually costs.
The third is the most badly estimated. Two addresses six hundred metres apart can take twenty minutes if you have to go round a main road, cross a market, or find somewhere to park twice.
The consequence is that a round optimised on distance alone is almost always worse than one an experienced driver would build by hand. That is not an argument against tools: it is an argument for giving them the three missing pieces of information.
And those three exist: they are in the head of the driver who has covered that sector for six months. The work described here consists largely of writing them somewhere, so that they survive their absence and their departure.
The order: what really decides the sequence
A round’s sequence is decided by four constraints, and distance is the last of the four.
The first is recipients’ opening hours. A shop delivered before it opens is a guaranteed failure; an administration delivered after three o’clock likewise. Those hours are known and are rarely in the system.
The second is traffic, which is predictable at the scale of a city and a day. A sector crossed at eight and the same at five do not take the same time, and that difference is larger than most optimisation gains.
The third is the constraints of the parcels themselves: what is collected on delivery, what is bulky, what has to be handed over in person. A heavy parcel at the back of the van imposes handling at every previous stop.
Only the fourth is distance. An order that respects the first three and travels five kilometres further is better than a geographically perfect one that presents two parcels at closed doors.
Density: one isolated address costs three
The cost of a delivery is not spread evenly. Twenty addresses in one district cost less time, together, than seven scattered over the same total distance.
The reason is that the real cost is not the journey but the stop: parking, getting out, going up, looking, calling, coming down, moving off. That sequence takes between four and twelve minutes and is almost independent of the distance travelled.
A commercial consequence follows that few carriers draw: an area where you have little volume costs you far more per parcel, and you generally charge the same price for it.
The answer is not necessarily to refuse those areas. It is to group them: delivering an outlying sector two days a week rather than every day multiplies its density by two or three, and it is the only lever available on that expense.
That decision has to be announced rather than endured. A recipient who knows their sector is delivered on Monday and Thursday arranges to be in; the same recipient delivered at random is out half the time, which returns the problem to where it started.
The call: two minutes before, not the day before
The advance call is the action that changes a round’s success most, and its timing matters more than its existence.
Called the day before, a recipient says yes and is not there. Called in the morning, they have gone to work. Called two to five minutes before arrival, they come down — because at that moment the question is no longer "will you be home tomorrow" but "I am downstairs".
The short call must contain three things and nothing else: who you are, that you are arriving in a few minutes, and a question about access. The third is the one people skip and it is what avoids five minutes lost looking for an entrance.
When nobody answers, the call has still served: it often triggers a call back within two minutes, and the driver can continue the sequence meanwhile rather than parking in front of a door.
The corollary is that the number has to be reachable. A wrong or illegible number turns this method into nothing, which is why the neighbouring article insists on the address and why the packing table in another article in this series verifies the number at sealing.
A failed address: three cases, and what gets written
A door that does not open covers three different situations, and telling them apart decides what happens next.
The recipient is out: the parcel is right, the address is right, you come back or leave it somewhere. The recipient refuses: the parcel reached them and they will not take it, which is commercial information for the sender. The address is wrong: nobody of that name, a building that does not exist, a district nobody can find.
The three look alike in a system where the driver picks a status from a list, in the rain, with the engine running. That is why a fourth piece of information is needed, written in full: what was observed.
"Building under works, entrance moved", "neighbour says he works until six", "number rings out, voicemail full". Those sentences cost fifteen seconds and they decide whether the second attempt has a chance.
They are also the only material that lets a cause be corrected. Twenty notes saying "district not covered by the street name" identify an addressing problem no status will ever surface.
The rule for the second attempt
The second attempt is the point at which a delivery operation makes or loses money, and it is almost always left to the judgement of the moment.
The rule that works is decided in the cold and fits in three lines. A second attempt is made if and only if the first produced usable information; it is made at a different hour from the first; and there is no third attempt without contact having been established with the recipient.
The first condition is the most important. Going back to the same place at the same hour knowing nothing more is pure expenditure: the probabilities are identical and so will the result be.
A different hour is a free adjustment and it is often enough. Somebody out at eleven is frequently in at five, and the reverse is true for a business.
A third attempt with no contact is never economically justified. At that stage you switch: a collection point, an agreed appointment, or a return to sender with the reason written — and that last is a decision, not a failure.
The collection point and the neighbour: two answers, two risks
Faced with a closed door there are two answers and they do not have the same standing.
A collection point is a clean answer: the parcel is handed to a third party who has a contract, premises and a record. It turns a failure into a controlled postponement, and its only limit is how far the recipient will travel.
Handing to a neighbour is a universal practice in this country and it settles many situations. It has a cost worth knowing: the carrier’s liability ends badly, the proof of delivery is weak, and the dispute that follows is between two people you do not know.
The rule that makes the practice tenable is simple: it is done with the recipient’s agreement obtained by phone at that moment, and the name of the person receiving is written down. Without both, it is a drop with no proof.
What a driver must never do is leave a parcel handed to nobody — outside a door, with an absent caretaker, in a stairwell. It is the only decision in this section that produces disputes nobody can arbitrate, and it is always taken to save two minutes.
The driver’s phone
The phone is the round’s main tool and it carries four things: the list, the numbers, the proof of delivery, and contact with base.
What it must not carry matters more: the full list of a day’s recipients is a file of personal data, and it has no business in a group conversation or on a personal device that goes home in the evening.
Proof of delivery deserves a precise requirement. A photograph taken in front of a closed door proves nothing; a photograph of the parcel handed over, or a signature, or failing that the written name of whoever received it, proves something. The difference between the two is three seconds.
Contact with base should be one channel and not four. A driver receiving instructions on three different channels while driving loses time and misses one — and the one they miss is always the one that mattered.
Finally, networks fail. A round has to be finishable with no connection: the printed morning list is a two-dinar redundancy that saves an entire day, and it is the same lesson as the run sheet in another article in this series.
Collecting money during the round
When payment happens on delivery, the driver carries cash, and that changes three things in how a day is organised.
The first is security, and it is handled through the frequency of deposits rather than through instructions. A driver who banks twice a day carries half as much; it is a trade between time and risk, and it should be made by management rather than by the person concerned.
The second is change, which is a real cause of failed delivery. An amount to collect against a large note with no change available produces a doorstep argument nobody can arbitrate, and sometimes a refusal.
The third is reconciliation. The total collected has to match the total expected, parcel by parcel, and that check is done the same day. Done weekly, it no longer allows anybody to know at which stop the gap occurred.
One substantive remark about this payment method: it shifts part of the sender’s work onto you and it is almost never priced accordingly. Collection time, change, evening counting and remittance are a service, not a modality.
The five-minute debrief
The end of a round is when the most useful information of the day is still available and nobody asks for it.
Five minutes are enough and cover three questions: which addresses failed and why in full, which stop took much longer than expected, and what fact about a sector deserves writing down.
The third is what builds the company’s capital. "Entrance to block C is round the back", "the Wednesday market blocks the street until two", "this shop closes between noon and three": durable facts, known to one person, and lost the day they leave.
Those notes go into a sheet per sector, not into a conversation. Information in a chat thread is unfindable three weeks later, and the stand-in for the day has no access to it.
This debrief is the only thing in this article that improves later rounds rather than today’s. It is also the first thing dropped when the day has run long, which is understandable and expensive.
What is counted, and why the billing unit hides the cost unit
Three things are counted per round, on the day’s sheet, and none needs a tool. Stops made, parcels handed over, and second attempts in the week.
The difference between the first two is the most useful figure in this whole trade and almost nobody keeps it. Fifty stops for forty parcels handed over means ten stops produced work and no revenue.
That is where the underlying problem sits, and it is structural rather than statistical: you are paid per parcel and you spend per attempt. A parcel delivered on the third pass earns once and costs three times, and nothing in your rate card records the difference.
The consequence is that a cost per delivery calculated the usual way — expenditure divided by parcels handed over — averages away precisely the variable this article is about. Two rounds with the same cost per parcel can have unrelated success rates, and it is the second that ruins you.
So count in attempts. Cost per attempt is the figure that describes your operation, and the ratio of attempts to parcels handed over is what says whether the problem is yours, the sender’s or the addressing. No published success rate replaces those two numbers, because they depend on the areas you serve and the senders you accept.
What we do, and what we refuse to do
What we do is writing and counting: the sheet per sector that collects durable facts, the second-attempt rule in three lines, and the count of stops against parcels handed over on the day’s sheet.
We do not build route optimisation software. An order calculated on distance alone is worse than an experienced driver’s, and until opening hours, traffic and parcel constraints are written down somewhere, a tool optimises the least important of the four constraints.
We will not put a day’s recipient list in a group conversation or on a personal device, even though that is what happens and even if asked. It is a file of personal data and it does not go home with somebody in the evening.
We will promise no first-attempt success rate, for the reason written in the previous section: that number depends on the areas served and the senders accepted, and a target set on it improves by refusing difficult areas rather than by delivering better.
Finally, the two most profitable actions are free: calling two minutes before arriving, and writing in full what was observed at a closed door. The first changes the day, the second changes the month.
Frequently asked questions
How should a round be ordered?
By four constraints, and distance is the last. Recipients’ opening hours first — a shop delivered before it opens is a guaranteed failure — then traffic, which is predictable and whose effect exceeds most optimisation gains, then the constraints of the parcels themselves. An order that travels five kilometres further and presents no parcel at a closed door is better.
When should the recipient be called?
Two to five minutes before arriving. Called the day before they say yes and are not there; called in the morning they have gone to work. At two minutes the question is no longer "will you be home tomorrow" but "I am downstairs". Three things in the call: who you are, that you are arriving, and a question about access — the third is the one people skip.
When should a second attempt be made?
Only if the first produced usable information, and at a different hour. Going back to the same place at the same hour knowing nothing more is pure expenditure: the probabilities are identical. And no third attempt without contact established — at that stage you switch to a collection point, an appointment, or a return with a written reason.
Can a parcel be left with a neighbour?
Yes, on two conditions: the recipient’s agreement obtained by phone at that moment, and the name of whoever receives it written down. Without both, it is a drop with no proof. What is never done is leaving a parcel handed to nobody — outside a door, in a stairwell — because the dispute that follows cannot be arbitrated by anybody.
What should be written when an address fails?
What was observed, in full, in addition to the status. "Building under works, entrance moved", "neighbour says he works until six", "number rings out". Fifteen seconds, and it decides whether the second attempt has a chance. It is also the only material that identifies a cause: twenty similar notes reveal an addressing problem no status will surface.
What first-attempt success rate should we aim for?
None from outside, and for a structural reason: you are paid per parcel and you spend per attempt. A parcel delivered on the third pass earns once and costs three times, and a cost per delivery calculated the usual way averages that variable away. Count stops against parcels handed over, and think in cost per attempt.
Where we come in
Stops made against parcels handed over, across one day: what separates them is what a second attempt actually costs you.
- We build the per-zone sheet that captures what the driver already knows.
- We reduce the second attempt to three lines readable at the wheel.
- We leave the two-minute call before arrival to your drivers.
Your customers’ addresses will go neither into a group chat nor onto a shared drive, and we will refuse to put them there.
Read next
AI in logistics: the list of abnormal parcels, before the customer calls
A network learns about its incidents by telephone. The parcel in trouble is the one that produced no event, and that is exactly what tracking cannot see.Logistics: you are graded by somebody who is not your customer
The merchant signs the contract. The recipient, who bought nothing from you, writes the review and refuses the parcel.The pilot that will never ship: the signs, from week two
A pilot that fails rarely fails at the end. It fails in the second week, quietly, and everybody carries on for three months.
Let us talk about your project
A free audit, no commitment: we look at your online presence and tell you what is holding it back.