Technology Forecasting Methods: 7 Techniques That Actually Predict Where Tech Is Heading

Deciding when to enter a technology, fund a programme or license an asset all rest on the same hard question: where is this field heading, and how fast? Technology forecasting methods are the structured techniques that answer it — turning patent filings, citations and expert judgement into a defensible view of a technology’s future rather than a hunch. This guide walks through the seven that matter most in practice, what each is good at, and how the best forecasts combine them.

What Technology Forecasting Methods Are For

Technology forecasting methods exist to convert scattered signals into a timing decision. They do not promise to name the winner; they estimate where a technology sits on its development path, how quickly it is moving, and which adjacent problems the next wave of work will attack.

The signals are real and measurable. Global patent applications passed 3.5 million for the first time in 2023, up 2.7% in a fourth straight year of growth, and roughly 2 million patents were granted worldwide, a 10.1% jump, according to WIPO’s World Intellectual Property Indicators 2024. Against that rising baseline, a field whose filings are merely flat is quietly losing ground — and only a method turns that observation into a decision.

Analysts group the techniques into two families. Exploratory methods start from where a technology is today and project forward. Normative methods start from a desired future and work back to what must happen for it to arrive. Strong forecasts almost always blend the two.

1. S-Curve Analysis: Reading the Life Cycle

The S-curve is the backbone of exploratory forecasting. Plotted as cumulative patent or publication activity over time, most technologies trace the same shape: a slow emergence phase, a steep growth phase as the approach proves out, a maturity bend as easy gains run out, and a saturation plateau as the field consolidates.

Its value is diagnostic. The same absolute filing count means opposite things at different points on the curve — rapid growth early signals opportunity, while the same growth flattening near the top signals a field about to consolidate. Fitting the curve to real filing histories and reading the inflection point, where growth stops accelerating, is the single most actionable output the method produces.

2. Trend Extrapolation and Filing Velocity

Trend extrapolation projects a measured trajectory forward — most usefully, the velocity of filings rather than their level. Reading how fast application volume is changing separates fields genuinely accelerating from those coasting on past momentum.

Velocity is best read in relative terms: a field’s growth rate against the whole-portfolio baseline and against its own three-year trailing average. Acceleration above baseline is the clearest early sign a technology is entering its growth phase; deceleration below it, well before the absolute count peaks, is the earliest warning that a field is maturing.

3. TFDEA: Forecasting the Performance Frontier

Technology Forecasting using Data Envelopment Analysis (TFDEA) tracks the rate at which a technology’s performance frontier advances across successive product generations. Instead of counting filings, it measures how much better each generation is on the attributes that matter, then projects when a target level of performance will be reached.

It is the method of choice when performance, not activity, is the question — how quickly battery energy density, chip transistor counts or sensor resolution are likely to improve. It complements the filing-based methods by forecasting capability rather than interest.

4. Bibliometric and Patent Analysis

Bibliometric analysis mines patents and scientific literature for structure — who is filing, who cites whom, and where activity is geographically concentrating. Citation networks in particular expose which clusters a field is organising around and which are being abandoned.

Peer-reviewed work has long shown the value of folding these signals into forecasting — see Daim and colleagues in Technological Forecasting & Social Change, who combine bibliometrics and patent analysis with growth curves, analogies and scenarios. The key advantage is timing: literature output often moves ahead of patent filings, making it a leading indicator of an emerging sub-field.

• New-entrant velocity — assignees filing in an area for the first time signal fresh capital arriving
• Citation bridging — patents connecting two previously separate clusters often mark a hybrid sub-field being born
• Classification drift — a rising share of filings under newer, finer codes shows examiners recognising a distinct area

5. Delphi and Expert Elicitation

Delphi is the workhorse normative method. A panel of domain experts answers structured questions over several anonymous rounds, seeing the group’s responses between rounds and revising toward a calibrated consensus. It is invaluable where data is thin or where the question is about desirability and feasibility rather than raw trajectory.

Its weakness is that experts share blind spots, so modern practice uses Delphi to calibrate data-driven forecasts rather than replace them — reconciling what the filing curves imply with what the people building the technology actually expect.

6. Scenario Planning and Technology Roadmaps

Where a single point forecast is fragile, scenario planning lays out several plausible futures and the timing signals that would distinguish them, so a strategy can be stress-tested against each. Technology roadmaps then sequence expected developments over time, typically built from a multi-stage Delphi survey combined with patent analysis.

Applied patent-analysis roadmapping frameworks — such as the one published in MDPI’s Sustainability — show how filing data anchors a roadmap in evidence instead of opinion. These methods are strongest at the strategy layer, translating a forecast into a plan leadership can act on.

7. Emerging-Field Detection in Practice

The highest-value forecasts flag a sub-field before its curve steepens. Artificial intelligence is the textbook case: filings at the EPO grew from 49 applications in 2015 to 2,243 in 2024 — a roughly twelve-fold rise, per the EPO Patent Index 2024. Teams watching new-entrant and citation signals saw that curve steepen years before the headline numbers made it obvious.

The same index shows why aggregate totals mislead. Overall European filings were flat in 2024 (-0.1%), yet computer technology became the largest field at 16,815 applications, energy-related electrical machinery grew 8.9% to 16,142, and biotechnology rose 5.4% to 8,479. A forecast built on sub-field velocity, not the top-line number, is what separates the accelerating fields from the coasting ones.

How to Choose and Combine Technology Forecasting Methods

No single technique is sufficient. The practical rule is to match the method to the question and then triangulate. Use S-curve analysis and filing velocity to place a technology and gauge momentum; TFDEA when performance improvement is the question; bibliometrics to detect emerging sub-fields; and Delphi to calibrate the data against expert judgement — then wrap the result in scenarios or a roadmap for decision-makers.

That triangulation is exactly how a structured technology forecasting engagement works, and it feeds naturally into sourcing decisions — as our biotech technology scouting case study shows, where reading each modality’s maturity redirected the search toward fields still open enough to enter on favourable terms.

How PerspireIP Can Help

At PerspireIP, our team helps innovators and businesses protect what they build. Whether you need a patent or trademark search, prior-art analysis, or an IP strategy tailored to your goals, we can help. Contact our team to discuss your next step.

Frequently Asked Questions