Script for Yusuf and Naeem

1. Opening

Zubair, 2 min. On screen: client site, Start.

Zubair: Thank you both for making the time, and for giving us the run of the bakery this week. On the 2nd and 3rd you told me what you wanted: three numbers every morning, production that follows sales so stale bread doesn't reach customers, what goes into the mixer checked against what comes out, and the bakery taking less of your time than it does now.

Zubair: This is how we'll use the hour. Mo will show you the site for a minute, then we go through the slides, and we finish with a couple of decisions we need from you.

2. The site

Mo, 1 min. Scroll Start slowly, then click through the tabs without stopping to read.

Mo: This is the site we built for you. Everything we found this week is in here, with the figures and how we know each one. Every tab goes deeper: how the bakery runs step by step, inputs against output, the morning report, the findings, a page for accounts, and a What if page where you can move the numbers yourselves.

Mo: You'll have all of this to sit with afterwards, so we won't read any of it now. The slides are the short version.

Switch to the slides tab.

3. The answer

Mo, 2 min. Slide 1, then slide 2.

Mo: Before anything else, we saw no carelessness on the floor. People write a lot down, and they write it carefully.

Mo: The problem is in how the figures are made. Your records agree with each other, because most of them are written from the standard rather than counted. So they can't show you where bread or time is lost. The fix is small. Count at two points on the sheet you already use, and put tomorrow's bake in front of Abu Bakr before the stores voucher is written.

4. Findings

Both, 15 min.

Slide 3. Where you are today (Mo, one line). About ZMW 2.9 million in sales a month, three quarters of it Deluxe, around 16 or 17 batches a day. And every van load is written in three books.

Slide 4. How the figures are made (Mo). A Deluxe batch is written as 216 loaves and 122 kg of flour almost every time. After the table, nothing counts the loaves again until the van. So the sheet tells you the standard, and it can't tell you what actually came out.

Slide 5. One batch, 5 October (Zubair). I watched batch 8 come out four short. The supervisor wrote 212 on the line, which is right. The total at the bottom was still written as 16 times 216. Accounts' Excel adds up the lines, so the books kept the four loaves. That's the pattern in one batch.

Slide 6. Damages (Mo, stop here). This is the one loss your books can see. About ZMW 51,000 a month at selling price, or about 30,000 in ingredients and bags. On average 35 Deluxe loaves a day damaged in the bakery and 27 returned damaged. Until this week each entry was a number with no stage and no cause. On 7 October, for the first time, the note was split: 21 in production, 9 at the slicer, 22 on sales. That split is what we want to keep.

Slide 7. Little goes missing elsewhere (Zubair, one line). The batches I counted came out full, 647 of 648 across three batches on Thursday, and your stock book closes every day. So we don't think much goes missing outside the damages line.

Slide 8. Tomorrow's bake (Mo, stop here). The bake is set the same morning, in a supervisor's diary and along the top of the stores voucher. Abu Bakr doesn't see it before it drives the requisition. Safi plans 25 batches every day to cover the morning bake. On Tuesday 18 were baked. Nobody checks the plan against the day.

Slide 9. Weekday pattern (Mo, one line). Sales don't swing much by weekday, so a clever forecast won't save much. A plan someone checks before flour is drawn is worth more, about 5 to 35 loaves a day.

Slide 10. Good news (Mo, one line). In 13 weeks no bread two days old was in stock overnight. Yesterday's bake carries over on Wednesdays to Fridays, so that's where the bake sheet starts.

Slide 11. Slicing (Zubair, stop here). We timed it at 41 loaves in 10 minutes on one bay, about 250 an hour. That's the bottom of what these machines are rated for, and by day it's usually one person. On an average day one bay would need about 14 and a half hours to get through the bread. So the trolleys queue. On Thursday all 13 were full and waiting in the cooler, and warm bread was being double stacked. That's the first link we've seen between slicing and damage, and it hasn't been counted yet.

Slide 12. Proof (Zubair, one line). Deluxe comes out of the proofer at 30 to 35 minutes. Your supervisors' cards say 40 to 45, and your own ideal is 46 to 50. The proofers have no timer.

Slide 13. Flour (Zubair, stop here). On Tuesday we were told equal thirds of Royal, Nyati and Nyimba. On Thursday half Royal was issued. The recipe needs at least half Royal. People gave us different splits, and the voucher only shows bags, so nobody can see if the mix changes.

Slide 14. Other products (Mo, one line). We went deep on Deluxe this week. Your records hold the same detail for the other six lines, and this is what they show. Cena is next in size.

5. Recommendation

Mo, 8 min.

Slide 15. One trial and three counts, all on the sheet people already use. Nothing new to learn, and anyone named can stop their part.

Slide 16. The bake sheet. By 08:00 each morning we'd send the batches to bake for each of the seven products, worked out from what sold on the same weekday over the last four weeks. Abu Bakr checks it and approves it, and the supervisor writes the voucher from it. It's an option. We'd suggest two weeks from 19 October, and Yusuf can stop it at any time.

Slide 17. Three counts. From 12 October: count good loaves after slicing on three Deluxe batches a day, keep the damage split by stage, and add up the batch lines for the day's total instead of writing batches times 216. And one food-safety step, dating each milk drum on arrival.

Slide 18. Who does what. No one has more than three things to do. Yusuf's part is a 20-minute review with Abu Bakr once a week.

Slide 19. When. Counts from the 12th, the trial from the 19th, and a review together on Monday 2 November, with the results in front of us.

6. Decisions

Both, 10 min. Slide 20.

Mo: So the decision for today is whether you'd like to try the bake sheet for two weeks. If yes, there are a few things we need from you.

1. Mo: Would you like to start on the 19th or the 20th?
2. Zubair: Should the sheet be printed and signed, or sent on WhatsApp to Abu Bakr and Safi?
3. Mo: Who should get the morning report by email?
4. Zubair: Yusuf, could you send the Excel from your timing project, even if it isn't finished? We'd like to put your times next to ours.

Write the answers down as they come.

7. Phase 2

Mo, 8 min. Switch to the site, Next.

Mo: This week we went deep on Deluxe. There's a lot more in your records, and these are the areas the week points to: handover and training, costs and margins by product, the monthly page accounts asked for, same-day figures, weekly damage by cause, and production against capacity.

Mo: Where would you want us to look next?

Let them talk. Write their pick on the page with them. If they raise Lee Yeast or Kafue Plastics, say the same method can move across when they want it. No price today.

8. Close

Zubair, 3 min.

Zubair: At three we're showing this to the wider team. Is there anything here you'd rather we kept out for now? For example the sales and margin figures, or damages in kwacha?

Zubair: Otherwise we'll run it the same way as this morning. Thank you both.

Likely questions

• "Is the 51,000 money we're losing?" It's at selling price. In ingredients and bags it's about 30,000 a month. Either way, it's the one loss the books can see.
• "Is someone stealing?" We found nothing that points to it. Little goes missing outside the damages line.
• "Can your forecast beat what we do now?" Not by much, honestly. Our estimate does about as well as last week's same day. The gain comes from following the weekday and having Abu Bakr check the plan.
• "How sure are you about slicing?" It rests on short timings on one bay. A full day's timing comes before the trial.
• "Who makes the bake sheet after the trial?" That's part of the phase 2 conversation, after the review on 2 November.
• "What would phase 2 cost?" Once we've agreed what you want first, we'll put it in writing.