Customer case study: Harlow Fresh

Harlow Fresh's 62 vans now drive 1,488 fewer miles a day

A regional grocer with three depots measured Routewise against the same twelve weeks a year earlier, using fuel-card receipts instead of our dashboard. Fleet miles fell 18.3%. Driver hours fell 6.4%, and on-time delivery barely moved.

140.6 miBefore: postcode zones, fixed order 109.4 miAfter: Routewise sequence
BeforeAfter
Van 14 out of Ashworth depot on one Tuesday: the same 24 drops, the same van, 31.2 fewer miles. One of the larger gains in the fleet; the average was 24 miles a van.

The same twelve weeks, one year apart

Weeks 19 to 30 of this year, after all three depots had switched, against weeks 19 to 30 of last year. Same vans, same depots, same delivery windows. Orders grew 4.4% in between, which is why miles per delivery fell further than total miles.

Three rows went nowhere or the wrong way. We left them in the table.

Daily averages across 72 delivery days in each period. Numbers in blue link to the method notes.
MeasureLast yearThis yearChange
Moved
Fleet miles per day8,1226,634−18.3%
Miles per van per day131.0107.0−24.0 mi
Miles per delivery22.912.28−21.8%
Diesel, gallons per day3452371−17.9%
Driver hours per van per day9.48.8−6.4%
Context
Orders per day2,7902,914+4.4%
Did not move, or got worse
Delivered inside the 1-hour window94.1%94.6%+0.5 pts
within noise
Failed deliveries (nobody home)1.8%1.8%0.0 pts
Miles per gallon17.9717.88−0.5%

Over a 312-day delivery year that comes to 464,000 fewer miles, 25,300 fewer gallons, $98,600 of diesel at Harlow's $3.90 contract price, and 257 tonnes less tailpipe CO2.4 Routewise costs Harlow $29,016 a year, so the fuel saving alone nets about $69,500.6

Routes drawn for nine stores, driven for thirty-one

Harlow Fresh started home delivery in 2014 with nine stores and one depot. The dispatch team split the region into postcode zones and gave each van a fixed zone order: Hollins Cross, then Pelham, then Fenner, then Carver Falls. Inside a zone, drops went in the order the orders arrived.

That worked while a van had fifteen drops. By last year the chain had thirty-one stores, three depots and 45 drops a van. Vans crossed the river twice a shift because Carver Falls came last on the list, and drivers who knew a better order had to break the printed sheet to use it.

We weren't bad at routing. We were running a thirty-one-store business on a zone list written when we had nine.

Marisol Okafor, Head of Transport, Harlow Fresh

Routewise replaced the zone order with a sequence calculated each morning from that day's drops, the delivery windows customers picked, and each van's load. Dispatchers still assign vans to depots. Drivers can override any stop, and every override is logged with a reason, which became the most useful data of the rollout.

I asked for fuel-card receipts, not dashboard screenshots. That's the only reason my name is on the 18%.

Priya Raman, Finance Business Partner, Harlow Fresh

Miles per delivery, week by week

The rollout made things worse before it made them better. Week 16 was the most expensive week of the year.

Miles per delivery, weeks 13 to 30Last year stays near 2.91 miles per delivery throughout. This year matches it until week 15, rises to 3.04 in week 16 when rollout starts, then falls to an average of 2.28 across weeks 19 to 30.rollout2.02.22.42.62.83.03.213151719212325272930week of yearLast year, avg 2.91This year, avg 2.28Wk 16: Ashworth goes first.Drivers override 31% ofsuggested stop orders.Wk 18: Millbrook, thelast depot, switches.Wk 24: Route 9 bridgeshut for four days.Wk 27: heatwave,orders up 9%, moredrops per street.
Blue: this year. Grey dashed: the same weeks last year. Shaded: rollout weeks 16 to 18, excluded from the comparison.1 Bracket: the twelve weeks compared in the results table.

What routing software did not fix

Shorter routes are one problem. These are three others Harlow hoped would improve alongside it. Two didn't, and one improved far less than the miles did.

  • Driver hours

    −6.4%9.4 to 8.8 hours per van
    grey last year, blue this year

    Miles fell 18%, hours fell 6%. Most of a driver's shift is spent at the door, carrying crates up stairs, and loading at the depot, and none of that is a routing problem. Harlow pays drivers by the shift, so the saved time went into taking two more drops per van.

  • On-time delivery

    +0.5 pts94.1% to 94.6%
    grey last year, blue this year

    Inside the week-to-week wobble of about ±0.9 points, so we count it as no change. Late deliveries at Harlow mostly come from late loading at the depot, which shorter routes don't touch.

  • Failed deliveries

    0.0 pts1.8% both years
    grey last year, blue this year

    A customer who isn't home is not home on a short route either. Harlow is testing a text message 20 minutes before arrival, which is a separate project and not counted here.

First fortnight it sent me over the Fenner level crossing at quarter past three, right when the school buses queue. I overrode it every single day until they let us block that crossing between three and four.

Dean Whitlock, driver at Ashworth depot for eleven years. Overrides in week 16: 31% of stops fleet-wide. By week 22: 6%.

How the numbers were measured

Case studies usually leave this part out. These are the choices that shaped every figure above, including the ones that flatter us.

  1. Comparison periodWeeks 19 to 30 this year against weeks 19 to 30 last year, so both periods cover the same school holidays and weather season. Rollout weeks 16 to 18, when miles went up, are excluded from the comparison and shown in the chart instead.
  2. Order growthOrders rose 4.4%. More drops in the same streets shortens the distance between them whatever software you use, so part of the miles-per-delivery gain would have happened anyway. Total fleet miles, which fell while orders grew, is the more conservative figure, and the one we lead with.
  3. Fuel is measured, not calculatedGallons come from Harlow's fuel-card transactions. Fuel economy dipped slightly (17.97 to 17.88 mpg), because shorter routes are a little more stop-start, so fuel fell 17.9% where miles fell 18.3%.
  4. EmissionsTailpipe CO2 at 10.18 kg per gallon of diesel, the US EPA factor. Refrigeration units run on separate fuel that was not metered, so they are not counted either way.
  5. What was left outHarlow's 11 refrigerated trucks that restock stores were not part of the trial. Vehicle maintenance probably fell with mileage, but nobody measured it, so we have not claimed it.
  6. Annual figuresDaily differences multiplied by 312 delivery days, at $3.90 a gallon, Harlow's average contract price over the period. If diesel prices or order volumes change, so does the saving. The licence cost is 62 vans at $39 a month.
  7. Nobody was paid for thisHarlow Fresh received no discount for taking part, and could have withdrawn the case study after seeing the numbers.

Harlow is one of our better results. Their starting point, a fixed zone order drawn by hand, is the least efficient way to route a fleet. If you already use routing software, expect a much smaller gain; the calculator below assumes so.

What would this mean for your fleet?

A deliberately cautious range, set below Harlow's result. Put in your own figures; nothing you type here is sent anywhere.

Units

We assume an 8 to 16% cut in miles.

$0 to $0

estimated fuel saving per year, before the licence

Fewer miles a year
0
Fewer gallons a year
0
Less tailpipe CO2
0
Routewise licence ($39 per van a month)
0
Net after the licence
0

Harlow Fresh's own figures in this calculator give $43,900 to $87,800. Their measured saving was $98,600, above the top of the range, and we still won't quote it to you as an estimate.

Fuel only. Driver time is left out because Harlow's hours fell just 6% and most fleets pay by the shift. CO2 uses 10.18 kg per US gallon or 2.68 kg per litre of diesel. Licence price for fleets up to 200 vans; larger fleets are quoted separately.

See your own routes replayed, not Harlow's

Send us one past week of delivery data and we'll run it through Routewise before we speak.

  1. You fill in the formA routing analyst replies within one working day with a short CSV template: stop address, time window, and van.
  2. You send one week of past dropsAny week. We delete the file after the call unless you ask us to keep it.
  3. A 40-minute call on your numbersYour old miles beside our sequence for the same stops. If the gap is small, we'll tell you so.

Rather talk first? fleets@routewise.example or +1 555 0142.

We use these details to arrange the demo and send the CSV template. No newsletter unless you ask for one.