The Shape of SaaS: A Morphological Operating System for Revenue Businesses
This is a hypothesis I've been working on, not a finished framework. I'm publishing it at the stage where the underlying claim is clear enough to argue, but the encoding still needs pressure-testing against real companies. Treat the rest of this piece as a sketch I'm thinking out loud against — agreement, disagreement and edge cases all welcome.
The starting point is something I keep noticing in operating reviews. You can look at thirty metrics and still not know what kind of business you're running. Not because the data is wrong. Because thirty numbers, however accurate, don't have a shape. They describe dimensions of a system without showing you the system itself. The best operators compensate for this by carrying a mental model they've built from years of pattern recognition. Everyone else stares at dashboards.
What follows is an argument that this mental model can be made explicit — that if you normalise six axes of operating health and plot them radially, a SaaS company stops being a spreadsheet and becomes a continuous geometric form. The form has properties you can read directly: it can be large or small, balanced or asymmetric, smooth or pinched, hollow or full. Those properties map back to how the business is actually performing in ways individual metrics can't. The shape is, in the most literal sense, a candidate visual language for reading a recurring revenue business.
I'm calling the working version a Morphological Operating System. It may not survive contact with real data. But I think the underlying intuition — businesses are systems, and systems have shapes — is sound enough to write down and test in public.
Why dashboards mislead more than they reveal
The problem with dashboards is not that they show too little. It's that they show the right numbers in the wrong structure. Each metric is a measurement of one dimension of the business, captured at a moment in time, presented without reference to the dimensions adjacent to it. You see the trees. You don't see the forest fire starting at the edge.
Jason Lemkin at SaaStr identified one version of this in what he calls the NRR Zombie: a SaaS company with stable 90–100% net revenue retention, sufficient cash and no obvious red flags, that has quietly lost the ability to acquire new customers. The dashboard shows health. The system is failing. A geometric reading would show it immediately — elongated toward retention, collapsing at acquisition, the silhouette of a business living off its past.
Tom Tunguz has documented a related trap. , the number most operators anchor on, assumes 100% retention — which is always an optimism. Hypergrowth masks this because new acquisition floods in before the churn cohorts mature. But the faster a company grows, the more it concentrates new customers at the highest-churn point of their lifecycle. The curve looks right until, suddenly, it doesn't. By the time the metric moves, the structural problem is months old.
Then there is the expansion illusion: above 120%, driven by upsells and seat growth among existing accounts, while new logo acquisition is quietly stalling. The net number looks strong. But expansion revenue is not the same as acquiring new customers who chose you over the alternative. It compounds differently. It is more fragile. A business running on expansion NRR with weak new business acquisition has a specific imbalance — and that imbalance should be readable before the trajectory bends.
These are not data failures. They are structural ones. The metrics are accurate. They simply don't show what they don't show.
The geometry of a system
Where I've ended up is this: SaaS businesses, viewed as systems, produce recurring imbalances. The same configurations of strength and weakness appear again and again across companies at different sizes, in different verticals, at different stages. The hypothesis behind a Morphological Operating System is that those imbalances are geometric — that a small set of axes, plotted carefully, captures more about the structure of a business than any longer list of metrics.
Radar charts have existed for decades and the idea of plotting metrics radially is not new. What I think is interesting is the move from polygon to continuous form. Six points connected with straight lines produces a hexagon, and a hexagon's geometry is dominated by the act of connecting points — its shape is mostly an artefact of the visualisation, not of the business. Smoothing those points into an organic curve gives the shape back to the data. The result is closer to reading an or a weather front than a tool's spider chart.
The six axes
I've ended up on six axes because they correspond to the structural moments of a recurring revenue engine — not because they're the only metrics that matter, but because together they describe the shape of the system. Each axis is a category, fed by one or more underlying operating metrics:
- Acquisition — the volume and quality of new revenue entering the system.
- Activation — the proportion of new users who reach the defined value moment.
- Retention — gross revenue retention, the core stickiness of the business stripped of expansion noise.
- Expansion — net revenue retention, the expansion-adjusted trajectory of existing accounts.
- Efficiency — how cheaply growth is being purchased. Inputs include and , the latter a concept introduced and benchmarked by David Sacks, where below 1 is excellent and above 2 signals structural concern.
- Engagement — depth of product use and stickiness inside the customer base, typically proxied by .
Each axis is scored on a 0–1 scale, where 0 is critical weakness and 1 is benchmark-leading strength. The idea is that any underlying metric — measured in months, percentages, ratios, dollars — can be normalised against a benchmark distribution and collapsed to the same 0–1 health score. The six scores are then plotted at fixed angular positions around a circle, clockwise from the top in the order above, and connected with a Gaussian-weighted radial field. The result is an organic, continuous shape: larger when the business is strong, pinched where it is weak, asymmetric where it is imbalanced.
To make this concrete, the next interactive lets you sketch your own business as six 0–1 weights and read the shape it produces.
Inflows are starved. Even strong activation, retention and expansion only compound on customers who arrived; until acquisition opens, the rest of the system is over-provisioned for traffic that isn't there.
No significant asymmetry — the system is broadly balanced across all six axes.
Recurring failure modes
If the hypothesis is correct, the same imbalances will keep showing up across very different companies. I don't want to hand them proper names yet — naming is the part most likely to harden a sketch into an orthodoxy it can't yet earn — but I do want to call out the patterns I keep seeing. Each of the silhouettes below is a structural failure mode I've watched companies live through.
Full and even across all six axes.
Reference state — every axis is reinforcing the others.
One axis spikes outward; the rest collapse.
Growth purchased on a single dimension that the rest of the system can't sustain.
Two strong opposite axes with a deep waist between them.
Strength at both ends of the funnel hides a bottleneck in the middle.
One axis bulges; the others stay thin.
A headline number is masking weakness in the dimensions that should support it.
Wide outer ring, one axis collapsed inward.
Surface metrics look strong but a foundational axis (typically engagement) is empty.
Strong and weak axes alternate around the circle.
A collection of disconnected bets rather than a coherent system. Imbalance is the risk.
The point of these examples is not to be exhaustive. It's to argue that there are a small number of recurring structural failure modes — not dozens — and that they are visually distinguishable as soon as you can read the shape. A balanced shape is the reference. Everything else is some specific way of being out of balance.
Translating metrics into axis scores
The six-axis interactive above takes 0–1 weights as inputs. The natural question is how you arrive at those weights from your own operating data. The piece below is a working aid: for each axis, it lists the metrics I'd reach for first, what a healthy reading looks like and what a weak one looks like, and lets you set your own score with the suggested ranges in view.
Treat this as a guide, not a calculator. A 0.6 on Retention from a company doesn't mean the same thing as a 0.6 from a public consumer SaaS. The benchmarks below are general; you'll know your category better than the defaults.
How well new revenue is entering the system.
- New logo growthhealthy 30%+ YoY ·weak <10% YoY
- Pipeline coveragehealthy 3x quota ·weak <1.5x
- Win rate (qualified)healthy 25%+ ·weak <15%
How effectively new customers reach first value.
- Activation ratehealthy 40%+ to value moment ·weak <20%
- Time-to-valuehealthy under a week ·weak more than a month
- Onboarding completionhealthy 70%+ ·weak <40%
How well the system holds onto core revenue once it arrives.
- GRRhealthy 90%+ ·weak <80%
- Logo churn (annual)healthy <5% ·weak >15%
- Cohort survival at 12mhealthy 80%+ ·weak <60%
How much the existing base compounds without new logos.
- NRRhealthy 120%+ ·weak <100%
- Net upsell ARRhealthy growing each quarter ·weak flat or declining
- Account expansion ratehealthy 30%+ of accounts ·weak <10%
How cheaply the growth is being purchased.
- CAC paybackhealthy <12 months ·weak >24 months
- Burn multiplehealthy <1.0 ·weak >2.0
- Gross marginhealthy 75%+ ·weak <60%
How deeply customers are using the product they pay for.
- WAU/MAU ratiohealthy 50%+ ·weak <25%
- Feature breadth usedhealthy 3+ core features ·weak 1 feature
- Power user sharehealthy 20%+ of seats ·weak <5%
Once you have six numbers, you can take them back to the shape interactive and see what the geometry says. The diagnostic panel will surface the weakest axis and the largest asymmetry, with a short note on what each typically implies.
How to read a shape
Reading a shape takes less practice than it sounds. Three properties carry most of the meaning.
Size reflects overall system strength. A large shape across all six axes indicates a business firing consistently. Size without coherence, however, is misleading.
Coherence is how evenly the axes contribute. A shape that curves smoothly belongs to a system where the parts are reinforcing each other. A shape with sharp irregularities means the system is fighting itself — fixing conversion breaks retention, improving activation strains onboarding, expanding into new segments dilutes engagement among the core base.
Pinches and asymmetries are the most actionable outputs. A pinch tells you which single axis is most constraining the system overall — work upstream of it has dampened impact until it opens. An asymmetry between two axes tells you which side of the business — acquisition or retention, efficiency or engagement — is pulling things out of balance. Most operators, shown a shape of their business for the first time, recognise the pattern immediately. The shape confirms what intuition already knew. The difference is that intuition couldn't be written down, compared quarter on quarter, or shared with a board.
Shape over time
If a single shape is a snapshot, a sequence of shapes is a prognosis. This is the part of the hypothesis I'm most interested in but also least sure about — partly because it requires longitudinal data I don't have, and partly because it's where the framework would have to do real predictive work to earn its keep.
The intuition: structural problems in SaaS systems should appear geometrically before they appear numerically. A pinch developing on the activation axis across three consecutive months, while overall MRR still trends upward, is a signal individual dashboards would never surface — because each metric, viewed in isolation, still looks acceptable. The shape would show the compression forming. The inverse holds too: a team that has invested deliberately in onboarding quality should see the activation axis expand and smooth months before any revenue impact registers. The shape would validate the investment ahead of the lagging indicators. That, ultimately, is what a diagnostic language is supposed to do — tell you something useful before it is too late to act on it.
Whether this actually works at scale is the next question I want to spend time on.
SaaS analytics has been additive for two decades. More metrics, more dashboards, more tooling layered on top of existing tooling. The result is more noise around the same signal. The argument I'm making — tentatively — is that businesses are systems, that systems have shapes, and that a reading layer above the dashboards might let operators say more clearly what kind of business they're running and where the structural risk sits.
The number of axes it takes to get to that sentence is six. The time it takes to read the answer, once you can read shapes at all, is under ten seconds. Whether the rest of the framework holds up is what I'd like to find out.
More to read
The Build vs Buy Decision Nobody Makes Properly
Every generation of enterprise software forces the same question. Generative AI hasn't settled it. It's made it harder, stranger and far more consequential — especially once you ask what you're actually building or buying.
Designing for the Person You Are, Not the Person You Think You Should Be
Most products are designed for idealised users who behave exactly as intended. Real users pile clothes on chairs, drop keys on tables and quit onboarding flows. The clutter is not a failure. It is the design brief.
Backlog Politics (Part II): Govern the Backlog, Don't Just Manage It
The antidote to backlog politics isn't a better scoring formula — it's governance. Outcome metrics, a North Star and transparent prioritisation criteria won't remove the argument. They'll just put it on record, and that changes everything about how the argument gets made.