Most technology forecasts are read for one number. Analysts scan for the compound annual growth rate, rank the sectors, and move on. A market growing at 23% goes on the shortlist; one growing at 7% does not.
That habit throws away the part of the forecast that describes the business you would actually be buying into. A single growth rate says how fast money arrives. It says nothing about what kind of money it is, who pays it, or whether next year’s revenue resembles this year’s.
The more useful question is whether the composition of revenue is changing. When a market’s growth rate falls while its mix moves toward higher-value, stickier work, the slowdown is not deterioration. It is the market improving and slowing at the same time, and the headline number conceals it completely.
Two Rates Are Published; Almost Nobody Subtracts Them
The raw material for this reading is already in most forecasts. Publishers typically state a historical CAGR for the preceding period and a projected CAGR for the forecast window, and the two are rarely compared.
Take Indonesia’s software development market, modelled from USD 5.62 billion in 2025 to USD 10.72 billion by 2031. Historical growth ran at 12.78% across 2020–2025; forecast growth moderates to 11.40% for 2026–2031. Ranked by rate, that is a market going quietly off the boil.
The report’s own account of why it moderates points the other way. Growth is expected to reflect complexity rather than project volume, with AI-enabled applications, security-by-design, data-residency controls and legacy integration supporting higher average contract values. Recurring modernisation and DevOps work should increase the share of multi-year contracts.
Those are not the characteristics of a decaying market. Longer contracts and higher contract values describe revenue that is harder to win and harder to lose. The rate fell while the quality of the revenue rose.
The Same Pattern, More Explicitly Measured
Vietnam’s warehouse automation market shows the mechanism with a number attached, which makes it the clearer case.
The market moves from USD 256 million in 2025 to USD 412 million by 2032, a forecast CAGR of 7.03%. Historical expansion was much faster, at an estimated 11.28% through 2020–2025. That is a substantial deceleration by any reading.
Underneath it, the composition shifts. Modelled software and control-system contribution rises from roughly 12.0% of market revenue in 2025 to 17.2% in 2032. The forecast also moves away from large greenfield hardware projects toward modular robotics, brownfield automation, execution software and recurring lifecycle services.
Normalised deployment volume, meanwhile, is expected to grow slightly faster than value, as component costs fall and modular systems reach mid-market operators. Volume up, price per unit down, software share up: three facts that a single 7.03% cannot express.
A vendor selling conveyors into that market faces a genuine slowdown. A vendor selling warehouse execution software faces a segment growing considerably faster than the market containing it. Both are described by the same headline rate.
Fast Markets Contain the Same Shift
This is not merely a way to feel better about slow markets. It applies equally where the headline is strong, and it changes what the strength means.
India’s SaaS market is forecast at 22.88% from USD 13 billion in 2025, against a 24.19% historical rate through 2020–2025. That is mild deceleration on a very high base, which alone would be unremarkable.
The composition tells you more. Value growth is expected to outpace normalised paid-seat expansion, with new revenue increasingly influenced by AI-enabled workflow automation, security, data intelligence and vertical SaaS.
Value outrunning seats means the market is monetising depth rather than headcount. For a vendor priced purely per seat, a large share of that forecast growth is not addressable. The market grows at 22.88% and that particular business does not.
Why the Aggregate Hides It
A market-level CAGR is a weighted average across segments moving at different speeds. Averaging is lossy, and what it discards is precisely the segment structure that determines whether a given company benefits.
This is why two firms competing inside the same forecast can experience opposite decades. The market rate is identical for both; the segment mix each sells into is not.
It also explains a common and expensive misreading. When a market decelerates because its cheapest, most commoditised tier is saturating while its higher-value tier accelerates, the aggregate falls as the opportunity improves. Ranking by aggregate rate points the wrong way.
Four Checks Before Trusting the Reading
The interpretation is easy to state and easy to over-apply. Four questions keep it honest.
Is a segment split actually published, or are you inferring one? The Vietnam software-share figures are modelled and stated outright. Where a report offers only a headline rate and qualitative commentary, mix shift is a hypothesis rather than a finding, and deserves to be labelled as one.
Are the historical and forecast periods constructed the same way? A revision to segment definitions, geographic scope or calculation basis can manufacture an apparent shift with no change in underlying demand. The Vietnam figures are explicitly stated on a 2025–2032 calculation basis; comparing them against a differently-built historical window measures the methodology instead of the market.
Does the shift favour what you actually sell? A rising software share is good news for a software vendor and a warning for a hardware integrator. The same sentence in the same report is a buy signal and a sell signal depending on who reads it.
Can you name the buyer and the contract? Higher average contract values should correspond to identifiable structural demand: regulated workloads, multi-year modernisation, compliance requirements. If the shift cannot be traced to a buyer with a reason to keep paying, treat it as unproven.
Where This Reasoning Breaks
Two failure modes deserve stating plainly.
The first is a market decelerating because demand is genuinely leaving rather than maturing. The arithmetic is identical to healthy mix shift, and only segment-level evidence separates them. If the fastest-growing segment is also the highest-value one, maturation is the better reading. If growth concentrates in the cheapest tier while premium segments stall, it is not.
The second is that every figure here is a model. A projection to 2031 or 2032 encodes assumptions about regulation, capital availability, talent supply and adoption rates that are not observations. Indonesia’s stated requirement for nine million digital talents by 2030 is exactly the kind of assumption that, if unmet, changes the forecast built on it. Treating a modelled trajectory as a structural fact is the most common way this analysis goes wrong.
What It Costs to Run
The check is a subtraction and a second look at the segment table. Both numbers usually sit in the same document, and most readers use only one of them.
What it buys is a defence against ranking markets by an average nobody actually sells into. The headline rate tells you how fast the money is arriving. The mix tells you whether you are the one being paid, and that is the question the shortlist was supposed to answer.
