Stay ahead of the curve
Get exclusive insights delivered directly to your inbox. Join thousands of tech leaders.
TL;DR
- Start moving when a challenger is improving faster than your core technology on the job your customers pay for. Margins, churn and lost deals move last.
- Compounding turns two improvement rates into a crossover date, so the rate gap tells you how much time you have and when to start.
- In the GetFocus Kodak backtest, digital separated from film in 1983. Film sales fell off a cliff in 2000: seventeen years of warning, while film still out-filed digital.
- Move in stages, with triggers set in advance: the core funds the transition, and every step is a decision rather than a debate.
Start moving when a challenger is improving faster than your core technology on the job your customers pay for, not when your margins start to slip. Margins, churn and lost deals are lagging signals: by the time they move, the transition has to be bought. The leading signal is the gap in improvement rates, and it shows up years ahead. In the GetFocus Kodak backtest, digital imaging separated from silver-halide film in 1983, while film still out-filed every digital technology. Film sales went off a cliff in 2000.
What are the signs that your core technology is being overtaken?
Every list of warning signs looks alike: customers start choosing the alternative, the challenger crosses a performance threshold, your margins come under pressure, engineers and suppliers move to the new approach, and competitors and investors pour money into it. Each of these is real. Most of them arrive late.
| Signal | What it actually tells you | When it moves |
|---|---|---|
| Lost deals and churn | Customers have already switched | Last |
| Margin pressure | The market has priced the shift in | Late |
| Engineers and suppliers moving | The ecosystem has placed its bet | Late |
| Competitor and investor money | Someone is convinced, not that the technology works | Any time, and often wrong |
| Challenger crosses a performance threshold | Where it stands today | Middle |
| A sustained gap in improvement rates | How fast the challenger is closing | First |
Watch all six. Trigger on the one that moves first.
Why do profitable incumbents move too late?
Because the reasons to stay are true. Kodak's leadership was not blind. Film quality was better for decades, film margins were extraordinary while digital's were thin, and Kodak did respond, with kiosks, Photo CD and an early digital SLR line. Each argument held at board level. None of them decided the outcome.
Kodak even saw it coming. After Sony's Mavica in 1981, an internal assessment concluded that digital would replace film in roughly a decade. The finding changed nothing. A forecast without a number to track is an opinion, and an opinion loses to a profitable P&L every quarter.
The rate is that number. The GetFocus Kodak backtest measured film at 23–26% a year against CCD's 30–90%. The rate does not show who is ahead. It shows who is closing.
Backtest · Imaging · 1980–2026
The fortress the filing chart kept confirming
Film out-filed every digital technology until 2010. The rate had separated digital from film in 1983.
What the industry filed · new patent families per year
What the rate read · improvement rate per year
Computed from patents public at each date · GetFocus backtest: film vs digital
Lighting repeated the pattern. In February 2007 GE launched a programme to double incandescent efficiency by 2010. That year LED was improving at 47% a year against the filament's 8.8%, more than five times faster. GE suspended the programme in December 2008 and sold its lighting business in 2020. The lighting backtest shows the gap was already four to one in 1990.
How can you tell when an emerging technology will overtake the one you use today?
Compare improvement rates on the same job, then let compounding do the arithmetic. MIT found, across 28 domains, that every technology improves exponentially, each at its own stable rate. Those rates span about twenty times: optical telecom improved 62% a year, milling machines 3%.
A stable rate makes the crossover computable. Take an incumbent improving 5% a year and a challenger improving 50%. The challenger gains about 43% a year on the incumbent (1.50 ÷ 1.05), so a challenger ten times worse today draws level in about six and a half years. Judge by today's performance and the incumbent looks safe for years. Judge by the rate and the crossover date is already on the calendar.
Independent research supports the arithmetic. Hindcasting 53 technologies, Farmer and Lafond (2016) showed how to estimate the probability that one technology will outperform another at a given point in the future.
GetFocus reads each technology's rate from the patent citation network, through cycle time and knowledge flow, validated at R² ≈ 0.7 on 150+ domains (how the method works). The same reading caught Kodak's second disruption: CMOS passed CCD on rate in 2000, five years before the filings agreed and fifteen before Sony ended CCD production.
A competitor just announced a big investment in the new technology. Should you follow?
Treat the announcement as a reason to check, not a reason to copy. Strategic fit, customer demand and your own economics all matter. Check the rate first, because money measures conviction, not progress.
The industry's money has backed the slower technology before. Automakers had spent more than $9 billion on fuel cells by 2014, and another decade of programmes and public funding took the total past $20 billion, while NMC batteries out-improved hydrogen from the first data point in 2001: 54% a year against 21%. GM paid about $50 million for rotary engine licences in 1970; at the first reading the rotary improved at 4.8% a year against the piston engine's 6.6%, and it never caught up. Following either bet meant joining the slower route.
If the competitor's technology is out-improving yours on the same job, its announcement is your trigger. If it isn't, the announcement is their bill. The mobility backtest shows the hydrogen bet year by year.
How do you move away without killing the cash cow?
Keep the core profitable, because it funds the move. Then shift in stages, each opened by a trigger you set in advance:
- Explore when the challenger's rate first pulls ahead: a small team, real customer tests, the skills you will need.
- Scale when the gap holds across several readings and the challenger wins a segment on total cost.
- Transition when staying costs more than moving: new development goes to the challenger, and the core runs for cash.
Work backwards from the crossover date. Subtract the years your migration takes and a buffer, and you have your latest start. Late usually has to be bought: Western Digital agreed to pay about $19 billion for SanDisk in 2015, fourteen years after the GetFocus storage backtest first flagged flash. (How to time the move.)
Put the triggers on the radar. GetFocus places every competing technology on one TRL × TIR radar with four verdicts: fund, monitor, phase out, ignore. Rates are recomputed as new patents publish, and an alert fires the moment a threshold you set is crossed. When your core technology slides toward phase out, the move follows a plan instead of a crisis.
This applies when a rival technology does the same job as your core and both leave a patent trail. It does not apply when the threat is a new business model built on the same technology, which no improvement rate measures.
FAQ
Should we stop investing in our core technology?
No. Fund it for cash and for the customers who will stay, and stop funding features that only extend your dependence on it. The core pays for the transition.
What is the earliest reliable sign that a technology will be overtaken?
A sustained gap in improvement rates on the same job. Digital separated from film in 1983, solid-state storage read 49.6% a year against hard drives' 29.2% in 2001, and LED read 32.5% against incandescent's 8.6% in 1990. Each signal came at least nine years before the industry's filings agreed.
How long do we have once a challenger is improving faster?
That depends on the size of the gap and how far behind the challenger starts, and compounding makes both computable. In the GetFocus backtests the warning was long: seventeen years from the Kodak signal to the collapse of film sales, fourteen from the storage signal to Western Digital's $19 billion purchase.
How do we avoid moving too early?
Stage the funding and set stop rules with each step. Early can be slowed down or scaled back; late usually has to be bought.
