Supply chain & project management Pakistan · 2026 A talk in five stops

One Crate
of mangoes

Follow a single crate of Sindhri from an orchard in interior Sindh to an export container at Karachi port. At every hop, a harder data problem is waiting on the road.

Every number in a supply chain is a compression, and every compression is a lie of a specific kind. This is a story about learning to read which lie you are looking at.
≈30%
of the crop is lost between farm and buyer2 — in a sector worth roughly a fifth of GDP. A number nobody actually weighed.
The route · farm → port

One crate. Four stops. At each one the same crate meets a different antagonist — and by the end, all four turn out to be the same problem wearing different clothes.

STOP 01 Farm gate · interior Sindh
Problem type — Measurement

The number that doesn't exist

Start with the headline everyone quotes: a third of the crop is lost between farm and buyer.2 Then ask the uncomfortable question — how does anyone know it is 30%? Nobody weighed it. There is no scanner at the farm gate, no timestamp, no record. The figure is an estimate laid over an informal economy where the real data lives in a broker's memory and a paper challan.

So the first hard problem is not analysis at all — it is measurement. In a formal economy the variable you would slice on already exists inside an ERP. Here it was never captured. That is the layer where Pakistani supply-chain companies actually get built: not better analytics on clean data, but the capture that makes data exist. Trukkr digitising trucking,3 Haball digitising B2B invoicing3 — measurement plays before they are anything else.

Farm-gate readout
Post-harvest loss30%+
Sector share of GDP~20%
Sensors at the farm gate0 — the variable was never recorded
STOP 02 The aggregator · the mandi
Problem type — Simpson's paradox7

The aggregate that lies

The crate reaches an aggregator buying from two supplier networks. On paper, network A passes quality at 90%, network B at 75%. The obvious move is to drop B. Make the room commit to it — out loud.

Now split by the one variable nobody logged: cold-chain availability on the route. B beats A on cold-chain routes and on routes with no cold chain. B only looked worse because it serves the hard rural terrain where no cold chain exists. The average measured B's terrain, not B's performance — and the buyer who consolidates to A just fired the better operator. At national scale, that single misjudgment is how smallholders get pushed out of formal supply.

Aggregate — the lie
Network A90%
Network B75%
Split by cold chain — the truth
Cold-chain routesB > A
No-cold-chain routesB > A
STOP 03 On the road · N-5 to Karachi
Problem type — Variance & tail risk

The average that hides the risk

Now the crate goes on a truck, and the question shifts from quality to time. Two carriers: one averages 10 days, one averages 14. The 10-day carrier looks better — until you show the spread. The fast one is 10 days give or take 8; the slow one is 14 give or take 1.

For a perishable crate with an export ship to catch, the reliable-but-slower carrier is worth more: you can plan around it, and the fast one strands your fruit more than a quarter of the time. Variance beats the mean. And with perishables it compounds — value does not just wait, it decays with every extra day, so the tail of that distribution is not a delay, it is a total loss. The number people optimise is the wrong number; the number that matters is not on the dashboard.

Carrier A — looks faster10 days ±8
Carrier B — actually plannable14 days ±1
0 days day 16 · ship cutoff 24
average transit typical spread past the ship — total loss
STOP 04 Exporter → retailer abroad
Problem type — The bullwhip effect8

The signal that distorts as it climbs

Zoom out to the whole chain. A small bump in real consumer demand becomes a larger order from the retailer, a larger one from the exporter, and a panic order back at the farm — each link padding against uncertainty in the one above it. This is the bullwhip effect,8 and it is fundamentally a data problem: no single party can see true demand, everyone reacts to the distorted echo of the party downstream, and the distortion amplifies as it travels up the chain.

The fix is not smarter forecasting at any one node — it is a shared view of the real signal, which no one in a fragmented, informal chain currently has. Name that for the founders in the room: the company that lets every link see the same demand truth is a real category, and it does not exist here yet.

Signal, amplified up the chain
Real consumer demand+5%
Order seen at the farm+40%
CauseEach link reacts to the echo, not the source
The turn

It was one problem the whole time.

A supply chain is a distributed system with no shared ledger. Every party sees only its own slice — so a small measurement gap, a lying aggregate, a hidden variance, and a bullwhip are not four separate faults. They are four faces of the same thing: running a system nobody can see whole, using compressed numbers that each hide a different part of the structure.

01 · Measurement
The number that doesn't exist
The confounder was never recorded.
02 · Aggregation
The aggregate that lies
The average measures the mix, not the thing.
03 · Variance
The average that hides the risk
The mean is optimised; the tail is ignored.
04 · Bullwhip
The signal that distorts
No shared view of the real demand.
Where the tool earns its place

The dashboard is a toy of the company.

Power BI, in forty minutes, is the instrument that reassembles the whole from the slices: group by the dimension, and watch the hidden structure surface. The toy you build on the slide is a miniature of the real venture in the market — because the durable Pakistani companies now are not flashy consumer apps. They are the unglamorous B2B tooling that instruments a workflow nobody had instrumented.5

The entry path is a supply-chain student's own career. You already understand the workflow. The gap is that you don't yet see the workflow as the product.4

A services firm already embedded in a client's process — logistics reconciliation, export compliance, SME operations — that moves from doing the work to shipping the tool built around it. That is where 2026 investment attention is concentrating.4

Reality check — don't sell a boom

Capital is real but selective: roughly $74.2M raised in 2025, nearly double 2024, still well below the $350M peak of 2021–22, and now flowing through hybrid equity-and-debt structures.1 Investors want cash flow, governance and a clear model — not hype.6 This is not a boom to chase. It is a quieter, more serious invitation: instrument something real.

You don't have to imagine the data. Every reversal here is a real, seeded dataset you can open in Power BI and rebuild by hand — the paradox, the tail, the bullwhip, all of it. The distance from this slide to a company is one dataset you already know how to read.
The discipline — so the tool isn't misused

You can always slice until something flips.

In a data-poor economy the temptation is worse, because the data is already thin — informal records, missing months, tiny samples. Slice thin data three ways and every cell becomes noise you can narrate into any story you wanted.

So the rule: disaggregate only on a variable you have a causal reason to trust as a confounder — and first ask whether that variable is even real, or just three lucky records you narrated into a pattern.

A BI tool amplifies your thinking, including your errors. The judgment stays with you.

Sources

References

Data — Pakistani market & economy · accessed Jul 2026
  1. invest2innovate (i2i) via Business Recorder — Pakistani startups raised ≈$74.2M in 2025, roughly double 2024, on a shift to hybrid equity-and-debt financing; still below the 2021–22 peak. brecorder.com · Feb 2026 · brecorder.com/news/40407474
  2. Startup Network Pakistan — Pakistan's agriculture is ~20% of GDP with post-harvest losses exceeding 30%; Tazah Technologies applies data analytics to demand, waste and logistics against that loss. startupnetwork.pk · Feb 2026 · startupnetwork.pk/10-pakistans-top-startups-you-must-know-about
  3. Daftarkhwan — Trukkr, a freight-tech platform, raised close to $10M (Yango Ventures) to streamline trucking logistics; Haball raised a $52M pre-Series A and has processed over $3B in payments digitising B2B invoicing and supply-chain operations. daftarkhwan.com · Feb 2026 · daftarkhwan.com/post/top-pakistani-startups-to-watch-in-2026
  4. Startup.pk — capital that survived the 2022–24 crash is concentrating into cross-border and B2B categories; the funded edge is services firms embedded in a client workflow that ship the tool built around it (services-to-SaaS). startup.pk · May 2026 · startup.pk/where-startup-capital-is-actually-going-in-pakistan-right-now
  5. Startups in Pakistan (mean.ceo), July 2026 — the durable opportunities are in software for supply chains, export workflows, compliance and SME operations; the "less glamorous" companies that become the durable ones. blog.mean.ceo · Jul 2026 · blog.mean.ceo/startups-pakistan-news-july-2026
  6. Startups in Pakistan (mean.ceo), June 2026 — a more selective market; investors now weight cash flow, governance and clear business models, with fintech, B2B software and logistics the strongest sectors. blog.mean.ceo · Jun 2026 · blog.mean.ceo/startups-pakistan-news-june-2026
Concept — the analytical ideas behind each stop
  1. Simpson's paradox (Stop 02). E. H. Simpson, "The Interpretation of Interaction in Contingency Tables," Journal of the Royal Statistical Society, Series B, vol. 13, 1951 — the foundational statement that an aggregate association can reverse once a confounding variable is separated out.
  2. The bullwhip effect (Stop 04). H. L. Lee, V. Padmanabhan & S. Whang, "The Bullwhip Effect in Supply Chains," Sloan Management Review, 1997 — with origins in J. W. Forrester, Industrial Dynamics, MIT Press, 1961 (demand distortion amplifying up a multi-stage chain).
The demo data — engineered, seeded, reproducible

The reversals in this talk run on a small seeded synthetic dataset, not on real firm records. The stop-level figures — network A vs B at 90% / 75%, the two carriers at 10 ±8 and 14 ±1 days — are engineered pedagogical illustrations, each reproducible by a single Power BI move: group by the dimension, and the hidden structure surfaces. The mandi's paradox is built twice over — once as a real cold-chain confounder that survives essentially every reseed, and once as a spurious slice with no mechanism, planted to be caught and discarded. Across a 3,000-seed sweep the real reversal holds 99.6% of the time; a clean fake one surfaces in barely 1.5%, and never favours a particular inspector — the signature of pure noise. That gap is the discipline lesson made literal. synthetic · seeded · reproducible in Power BI · 2026