Half-life is the time for half of something to decay or disappear. In physics, it describes radioactive isotopes: after one half-life, half the atoms have decayed; after two, a quarter remain; and so on. The same curve applies to anything that decays exponentially: knowledge that becomes obsolete, skills that atrophy, audiences that churn, and the value of information over time. The model gives you a single number that captures how fast something loses relevance. Short half-life means rapid decay; long half-life means persistence.
Strategy and operations are full of half-life dynamics. The half-life of a technical skill might be a few years as the stack changes. The half-life of a marketing channel's effectiveness can be months. The half-life of a competitive advantage varies by industry — some moats decay fast, others last decades. Customer retention curves often look like decay: a fixed fraction of the base leaves each period. The strategic question is always: what is the half-life of the asset or assumption I am relying on? If it is short, you must renew constantly or accept obsolescence. If it is long, you can invest with a longer horizon.
Half-life also applies to the impact of decisions. The effect of a reorg or a product launch may decay — half the impact in six months, half again in the next six. Planning and investment should match the half-life: don't build a five-year strategy on an asset with a one-year half-life. Don't treat a long-half-life asset (brand, relationship) as if it will vanish tomorrow.
Section 2
How to See It
Half-life shows up when a quantity or value decreases by a consistent proportion over equal time intervals. Look for decay curves, churn rates, and "how long until this is half as relevant?"
Business
You're seeing Half-life when a cohort of customers or users churns at a predictable rate — e.g. 5% per month. The "half-life" of the cohort is the time until half have left. Retention strategy is extending the half-life: better product, habit, or switching costs so that decay slows. Short half-life means you must acquire constantly to maintain the base.
Technology
You're seeing Half-life when the value of a technical skill or a codebase decays. A framework has a half-life — after a few years, half the knowledge may be obsolete. Documentation and systems have a half-life: they decay unless maintained. Roadmaps should account for the half-life of assumptions and dependencies.
Investing
You're seeing Half-life when an edge or information advantage decays. Insider information has a very short half-life. A structural moat may have a long half-life. Valuation should reflect how long the cash flows are likely to persist — i.e. the half-life of the advantage that generates them. Mean reversion is half-life thinking: excess returns decay.
Markets
You're seeing Half-life when the effectiveness of a channel or tactic decays. A new ad format works until saturation; its half-life might be months. A trend has a half-life. The first mover captures value while the half-life is long; late entrants face a decayed opportunity. Position for the half-life that matches your strategy.
Section 3
How to Use It
Decision filter
"Before relying on an asset, assumption, or advantage, estimate its half-life. How long until half the value has decayed? Match your investment horizon and renewal rate to the half-life. Short half-life assets need constant renewal; long half-life assets can be held and compounded. Don't confuse the two."
As a founder
Estimate the half-life of your key assets: team knowledge, product relevance, channel effectiveness, and competitive advantage. If your moat has a short half-life, you must invest heavily in renewal — R&D, learning, experimentation. If it has a long half-life, you can compound on it. Structure the company and roadmap so that renewal keeps pace with decay. The mistake is building as if a short-half-life asset will last forever.
As an investor
Price in half-life. How long will this company's advantage last? What is the half-life of the demand, the technology, or the regulatory window? A high multiple on a short-half-life business is dangerous. A reasonable multiple on a long-half-life business can be justified. Ask management how they think about the decay of their edge and how they plan to extend it.
As a decision-maker
When allocating time or capital, consider the half-life of the payoff. Training that decays in a year may not be worth a five-year commitment unless you renew. A decision whose impact decays quickly needs follow-up and iteration. A decision that creates long-half-life value (e.g. a relationship, a brand) deserves more upfront investment.
Common misapplication: Assuming half-life is fixed. Half-life can change. A product improvement can extend the half-life of retention. A new competitor can shorten the half-life of your advantage. Regulation can extend or shorten the half-life of a business model. Re-estimate as conditions change.
Second misapplication: Treating all decay as exponential. Some things decay linearly or in steps. Half-life is the right model when decay is proportional to what remains — the classic exponential decay curve. When decay is driven by one-time events (e.g. contract expiry), a different model may fit. Use half-life when the process is continuous and proportional.
Buffett seeks businesses with long half-lives: "our favourite holding period is forever." He prefers competitive advantages and demand that decay slowly — brands, regulated monopolies, essential products. He avoids businesses whose relevance has a short half-life (e.g. fast-changing tech where today's leader may be obsolete in a few years). Half-life is implicit in his margin-of-safety and durability-of-moat framework.
Grove's "only the paranoid survive" is a response to the short half-life of tech advantages. He argued that strategic inflection points can dramatically shorten the half-life of an incumbent's position. Intel had to keep reinventing to stay ahead of the decay curve. The lesson: in high-decay industries, renewal must be built into the culture and strategy.
Section 6
Visual Explanation
Half-life — Time for half to decay. Exponential: after 1 half-life 50% remains, after 2 half-lives 25%, etc. Match horizon and renewal to half-life.
Section 7
Connected Models
Half-life connects to compounding, time horizon, and decay. The models below either describe the opposite dynamic (compounding, exponential growth), the same math in another form (decay), or how to use time (time horizon, diminishing utility, Lindy).
Tension
Compounding
Compounding is growth that builds on itself; half-life is decay that compounds against you. Both are exponential. Compounding extends value over time; half-life shortens it. The best assets compound and have long half-lives; the worst decay fast. The two models are mirror images.
Tension
Exponential Growth
Exponential growth doubles in fixed periods; exponential decay halves in fixed periods (half-life). The same mathematical structure, opposite direction. Growth has a "doubling time"; decay has a half-life. Strategy: seek growth in what you own and long half-lives for what you depend on.
Reinforces
Time Horizon
Time horizon is the period over which you plan or invest. It should match half-life: don't plan on a 10-year horizon for an asset with a 2-year half-life unless you have a renewal strategy. Half-life tells you the natural horizon of the asset.
Leads-to
Diminishing [Utility](/mental-models/utility)
The value of an additional unit often decays with quantity or time — diminishing utility. Half-life is one way to quantify that decay over time: how long until the utility (or relevance) has halved? The two combine when you model how much something is "worth" over time.
Section 8
One Key Quote
"Our favourite holding period is forever."
— Warren Buffett
Forever only works if the asset has a very long half-life. Buffett's discipline is to buy businesses and advantages that decay slowly — so that the compounding can run and the half-life does not force a sale. Half-life is the hidden constraint behind "hold forever."
Section 9
Analyst's Take
Faster Than Normal — Editorial View
Estimate half-life for every key asset. How long until half the value has decayed? Skills, channels, moats, and information all have half-lives. Strategy that ignores decay assumes infinite half-life — a dangerous assumption in most markets.
Match horizon to half-life. Don't build a five-year plan on a one-year half-life advantage. Either plan for renewal (so effective half-life extends) or shorten the horizon. Mismatch causes overinvestment in decaying assets or underinvestment in renewing them.
Short half-life industries need built-in renewal. Tech, fashion, and fast-moving markets have short half-lives. The winners are those that institutionalise renewal — continuous learning, rapid iteration, and culture that accepts obsolescence. The losers are those that treat the current advantage as permanent.
Long half-life assets are undervalued when the market discounts short.Brand, trust, and relationship can have very long half-lives. If the market prices them as if they decay in a few years, there may be opportunity. The discipline is to estimate half-life honestly and price accordingly.
Section 10
Summary
Half-life is the time for half of a quantity to decay. It characterises exponential decay — of atoms, retention, relevance, or advantage. Strategy: estimate the half-life of what you depend on, match your horizon and renewal rate to it, and avoid building long-term plans on short-half-life assets. Long half-life allows compounding; short half-life demands renewal.
Grove on strategic inflection points and the short half-life of tech advantages. Renewal and paranoia as a response to decay.
Reinforces
Lindy Effect
The Lindy effect suggests that the future life expectancy of some things increases with how long they have already lasted. Long-lived things have long half-lives. Lindy is a heuristic for "this has a long half-life"; half-life is the explicit measure. Use Lindy to guess half-life for non-quantifiable assets (ideas, institutions).
Leads-to
[Entropy](/mental-models/entropy)
Entropy tends to increase; order decays toward disorder. Half-life is one way to quantify decay — how long until half is gone. In strategy, entropy is the natural erosion of advantage, knowledge, or relevance; half-life is the measure of how fast that erosion proceeds.
