One stack, many shapes
Every company runs on the same information stack, but the stream that hurts you first depends on how you make money. Sriram Chidambaram walks the archetypes (SaaS, platforms, D2C, manufacturing, deep tech) and the two streams almost everyone under-instruments.
Summary
- The information stack is universal, but the binding constraint is not: the stream that hurts you first depends on how your business makes money.
- SaaS lives or dies on revenue definitions, platforms on separating GMV from revenue, D2C and manufacturing on inventory and working capital, deep tech on keeping three clocks apart.
- Almost everyone under-instruments the people stream, and anyone selling a physical product has cash hiding in inventory.
I've made the case that every company runs on the same information stack: three data streams (sales, people, operations) climbing five layers (source, measurement, meaning, decision, confidence). That's true, and it's the point of the framework: the architecture is universal.
But universal doesn't mean identical. The stack is the same everywhere; what breaks first is not. The binding constraint (the stream that will hurt you soonest, and hardest) changes completely depending on what kind of business you are.
Same cockpit, different instruments. And knowing which of your instruments matters most, given your model, is most of the art. Let me walk the main archetypes, then two streams almost everyone under-instruments regardless of type.
SaaS: the sales stream, and the war over "revenue"
For a SaaS company, the binding work sits at the bottom of the sales stream: the Source and Measurement layers. It sounds mundane, but it's where SaaS businesses live or die in diligence: the definitions of bookings, billings, and recognised revenue, and the discipline to keep them straight.
A SaaS founder who can't cleanly bridge ARR to bookings to recognised revenue will lose a round they should have won, because subscription revenue is deceptively easy to overstate and every serious investor knows exactly where to press.
Above that, the metrics that matter are retention and : net revenue retention, gross margin, the true cost to acquire and keep a customer. SaaS makes the sales stream look simple because the product is intangible; that's the trap. The intangibility is exactly why the definitions have to be airtight.
Platforms and marketplaces: the two-sided truth
Platforms have a subtler version of the same problem. The binding challenge is measuring a two-sided business honestly. GMV is not revenue. looks stable until you decompose it. Supply-side and demand-side metrics each tell half a story, and the half you choose to lead with can quietly mislead, including yourself.
For a platform, the Source and Measurement layers have to hold the distinction between the volume flowing through the platform and the money the platform actually keeps. Founders who blur GMV and revenue in their own heads make bad allocation decisions long before an investor ever catches it. The information discipline here is refusing to flatter yourself with the bigger number.
D2C: the operations stream, cohort by cohort
For a direct-to-consumer brand, the centre of gravity shifts hard to the operations stream. The whole game is , worked out cohort by cohort: the real cost of acquiring a customer, fulfilling their order, and serving them over time, against what they actually spend.
A D2C business can be growing topline fast and destroying value on every order, and a weak information stack will hide that until the cash runs out. Two things dominate: fulfilment economics and inventory. Which brings me to the single most expensive blind spot in physical businesses.
Inventory: where cash goes to hide
If you sell a physical product, D2C or manufacturing, this is the deep dive that matters most, and it's the one I've watched sink otherwise healthy companies.
Here is the uncomfortable truth: a company can be profitable on paper and insolvent in practice, and the gap between the two is almost always hiding in inventory. Your P&L can show a healthy margin while your bank balance quietly tightens every month, because the money isn't lost, it's frozen, locked up in stock sitting in a warehouse that no report is clearly showing you.
I've seen founders celebrate record revenue while their cash position deteriorated, genuinely puzzled, because their information stack measured sales beautifully and inventory not at all. Inventory turns had collapsed, ageing stock was piling up, and working capital was silently draining, and none of it appeared in the numbers they looked at. Revenue was fine. The operations stream was blind.
Fixing this isn't exotic. It's giving the operations stream the same instrumentation the sales stream already has:
- Real-time visibility of what's in stock and what it's worth.
- Inventory turns and days-inventory-outstanding, tracked like the vital signs they are.
- A clear line of sight from a purchase commitment to the cash it consumes, before it becomes a surprise.
Inventory is where cash hides.
If you sell a physical thing and can't see your inventory clearly, you can't see your business.
Manufacturing: throughput, margin, and working capital
Manufacturing takes the inventory problem and adds a second one: turning throughput into margin. The binding work is the operations stream's ability to connect what the plant produces to what the business actually earns: cost per unit, , the true cost to serve, and where margin leaks between the factory floor and the invoice.
Manufacturers also carry the heaviest working-capital load of any archetype: cash tied up in raw materials, work in progress, and finished goods all at once, plus the timing gap between paying suppliers and collecting from customers.
For a manufacturer, an information stack that can't see working capital in real time isn't an inconvenience; it's an existential risk, because the business can be fundamentally sound and still run out of cash on a timing mismatch.
Deep tech: the three clocks that must never collide
Deep tech is the archetype where the information problem is most literal, and where getting it wrong is most lethal in a raise. A deep-tech company runs on three separate clocks that must never be allowed to collapse into one another:
- The milestone clock: technical and R&D progress.
- The bookings clock: contracts signed, often long before delivery.
- The recognised revenue clock: what's actually been earned.
Confuse them (and it's dangerously easy to, because the gaps between them can be measured in years) and you either mislead your investors or mislead yourself.
The Confidence layer is where this becomes existential. Deep-tech diligence is unusually brutal precisely because sophisticated investors know these three clocks drift apart, and they check. A deep-tech founder whose milestone story, bookings, and recognised revenue don't reconcile cleanly will be marked down not for the science, but for the sense that they don't have a grip on their own numbers.
The people stream: the one everyone forgets
Across every one of these archetypes, there's a stream almost nobody instruments properly, and it's not sales or operations. It's people.
Every founder instruments sales, because that's what everyone watches. Most get to operations eventually, usually after inventory or cash forces the issue. Almost no one instruments the people stream, until a key person resigns and it turns out the data had been signalling it for months, if anyone had been reading.
That's the tell. In most companies, attrition is discovered, not predicted. Someone hands in their notice and it's treated as a surprise, when tenure, engagement, comp position, and workload had been pointing the same way for two quarters.
The people stream, treated properly, is as rich a funnel as anything in sales: a lifecycle from hire to exit that can be measured, understood, and forecast. Instead it lives in a spreadsheet nobody trusts.
And it matters more than founders assume, because people are usually the largest cost and the actual engine of the business. The questions a mature people stream answers (what's our real output per person, what does an outcome actually cost us, which teams are about to break, who's about to leave) are not soft HR questions.
They're among the most important numbers in the company, and most companies can't answer a single one of them with data. Instrument the stream you find least interesting.
Your blind spot is almost always the stream you don't naturally love.
Same framework, different first move
So the stack is universal, but the entry point is not.
- SaaS: lock the revenue definitions.
- D2C: cohort contribution margin, and inventory.
- Manufacturing: working capital.
- Deep tech: separate the three clocks, and get them to reconcile.
And for almost everyone, there's a neglected people stream and, if they sell anything physical, a blind spot where inventory hides the cash.
This is exactly why there's no single template that fixes reporting, and why "best-practice" dashboards borrowed from another kind of company so often disappoint: they instrument the wrong stream for who you are. The right approach starts with a different question: given what kind of business I am, which stream will hurt me next, and is it instrumented well enough to see the pain coming?
Answer that honestly, start there, and climb the stack from your real weakest point rather than your most comfortable one. Same cockpit, different instruments, but every pilot has to know which dial, on their particular aircraft, is the one they cannot afford to fly blind.
How relevant and useful is this article for you?
The next one
Get what we publish next, by email.
Working notes on raising, borrowing, protecting, growing and structuring capital in India. One email a week at most, and you can leave any time.
We use your address only to send this. See our privacy policy.
We store your address to send you these emails and nothing else. See our privacy policy.
Related reading
