Most pages headed “take action” list things that are easy to do, not things that work. The difference has been measured.
The honest framing first. Individual behaviour is not how atmospheric warming gets solved. Global mean temperature in 2025 sat at approximately 1.43°C above pre-industrial levels, and atmospheric CO2 reached 425.6 ppm, a 53% increase over the pre-industrial figure of around 278 ppm. Quantities on that scale move when energy systems, industry and land use change, not when households do.
But “not sufficient” is not “not worth doing”. The actions available to a household differ from each other by an order of magnitude or more, and knowing which is which is the difference between climate action that registers and climate action that mainly feels like it.
What the research measures
The most useful ranking comes from Seth Wynes and Kimberly Nicholas, “The climate mitigation gap”, published in Environmental Research Letters in 2017 and associated with Lund University. They reviewed 148 scenarios from 39 sources and converted individual actions into one comparable unit: annual tonnes of CO2-equivalent avoided per person.
Three caveats belong in front of the numbers. They are averages for developed countries — a household on a coal-dominated grid and one on a hydro-and-wind grid get very different results from the same action. They describe emissions associated with a person’s consumption, not the change in the global total once markets respond: decline a flight and the seat usually sells to somebody else. And the study is several years old; changing grids and vehicle fleets have shifted the ranking without reversing it.
The paper’s central finding was a gap: none of the government guides the authors reviewed recommended the two highest-impact actions they had identified. The guides recommended what was comfortable to recommend.
The high-impact actions
Living car-free: 2.4 tonnes CO2-equivalent per year. The largest recurring action available to most people in high-income countries. It removes at once a manufacturing footprint amortised over the vehicle’s life, a continuous fuel demand, and a share of the road and parking infrastructure built to serve it, which is why the figure is so large. For most households the realistic version is car-light rather than car-free, capturing a fraction rather than the whole.
Avoiding one transatlantic return flight: 1.6 tonnes CO2-equivalent. Note the unit: one journey, not a year of them. Two long-haul returns annually exceeds the benefit of giving up a car entirely. Per hour of activity, flying is the most carbon-intensive thing most people ever do, and no substitute for jet kerosene exists at the volumes required — so these emissions are not on the electricity sector’s decarbonisation curve.
Eating a plant-based diet: 0.8 tonnes CO2-equivalent per year. Dominated by ruminant methane and by land converted to grow feed. Beef and lamb account for most of the total, so partial shifts capture much of the benefit: dropping red meat while keeping dairy and eggs lands far closer to the plant-based figure than to the starting point.
Home heating. In cold climates heating is the largest single household energy use, though the paper does not quantify it in the same units. A heat pump shifts that demand onto a decarbonising electricity grid and off a fuel that is not. How much it gains you depends on local generation, covered below.
The figure the study is known for. Wynes and Nicholas placed “having one fewer child” at the top of their ranking, at 58.6 tonnes CO2-equivalent per year for developed countries. That number is measured in a fundamentally different way from the others and belongs in its own category. It uses generational accounting: a parent is assigned a share of their descendants’ lifetime emissions across future generations, divided by the parent’s own remaining years. The method is contested — it assumes future emissions intensity resembles today’s, and the result shifts depending on where the boundary between generations is drawn. The resulting per-year figure is not comparable to a per-year figure for cycling to work. Family size is a personal decision shaped by considerations that have nothing to do with carbon accounting, and this site does not advise on it. We report the finding because it is central to the paper’s argument about which recommendations get made.
The actions that get recommended anyway
Recycling. Unplugging chargers. Refusing straws. Shorter showers. Lights off in empty rooms. These appear on nearly every list.
Comprehensive household recycling is the largest of them, and the comparison is instructive. A plant-based diet saves roughly four times more greenhouse gas emissions per year than comprehensive recycling; a single avoided transatlantic return flight saves roughly eight times more. Chargers left plugged in draw a fraction of a watt, and the annual total is measured in kilograms rather than tonnes.
None of that makes them worthless: recycling has resource and waste benefits independent of carbon, and water heating is genuinely energy-intensive. They are simply an order of magnitude smaller than the actions above.
The sharper problem is what small actions can do to large ones. Research on moral licensing finds that people who have recently done something they consider virtuous become more willing to relax elsewhere. In experiments published in Nature Climate Change, exposure to a green energy default reduced support for a carbon tax across six studies. The mechanism is not hypocrisy but the ordinary sense of having already contributed. Small actions are safest treated as habits rather than as the contribution.
The context that changes everything: your grid
Every electrified action inherits the carbon intensity of local electricity. The global average sat somewhere around 445 to 473 grams of CO2 per kilowatt-hour in 2024 depending on source and method, but the spread behind it is enormous: grids running largely on hydro, nuclear and wind come in below 100, while coal-dominated grids run several times higher.
The same heat pump on those two grids gives very different first-year results. It delivers roughly three to four units of heat per unit of electricity, so it still beats a gas boiler on most grids — but the margin is narrow where generation is coal-heavy and decisive where it is clean. Electric vehicles behave the same way: the advantage over petrol narrows on a dirty grid without disappearing.
The point most often missed is time. A heat pump or an electric vehicle lasts fifteen to twenty years, and the grid underneath it changes over that period. Utility-scale solar costs fell about 90% between 2010 and 2024, onshore wind about 70%. Equipment electrified today gets cleaner every year it runs; a gas boiler installed today emits exactly as much in 2040 as it does now. That asymmetry is why electrification usually wins across the life of the equipment even where it barely wins on day one. Most system operators publish grid carbon intensity, often hourly.
Where individual action hits a hard ceiling
Energy production accounts for roughly two-thirds of global greenhouse gas emissions, but the demand it serves is mostly not domestic. Iron and steel account for around 7% of global emissions on their own; cement roughly 3%; aviation and shipping around 2% each. Agriculture, forestry and land use together make up close to a fifth. Then there is the building stock already standing, most of which still will be in 2050.
The clearest framing comes from the International Energy Agency’s Net Zero by 2050 pathway. In it, around 55% of cumulative emissions reductions are linked to consumer choices — buying an electric vehicle, retrofitting a home, installing a heat pump. Behavioural changes such as replacing car trips with public transport, or foregoing a long-haul flight, provide around 4%.
Read them together. Consumer decisions matter enormously, but nearly all of that 55% depends on what is available to buy, what it costs, and whether the infrastructure exists. Behaviour alone, with technology and prices fixed, contributes a few percent.
What actually scales
Political participation, particularly local. Grid connections, planning consent for wind and solar, building codes and public transport funding are decided at municipal, regional or utility-regulator level, often in meetings almost nobody attends. One change to a building code applies to every structure built under it for decades.
Workplace and institutional decisions. Procurement, vehicle fleets, building energy, travel policy, electricity contracts. Anyone able to influence an employer’s energy contract is moving emissions orders of magnitude larger than their household total. Universities, hospitals and councils behave the same way.
Pension and savings allocation. Where long-term savings sit determines what gets financed. UK campaign research by Make My Money Matter with Aviva and Route2 estimated that moving an average £30,000 pension into a sustainable fund corresponds to around 19 tonnes of CO2 a year — a campaigner’s figure, assigning savers a share of investee-company emissions, and best read as indicative.
Talking about it. The weakest-sounding item here, with the clearest mechanism behind it. In Yale’s spring 2026 survey, 69% of Americans said they rarely (34%) or never (35%) discuss global warming with family and friends. Silence is self-reinforcing: people underestimate how much those around them care, which suppresses both the perceived expectation to act and the perceived support for policy. Conversation is not a substitute for the actions above. It is how they spread.
How to check a climate claim
Four questions worth asking of any organisation’s climate commitment.
Absolute or intensity? An absolute target commits to fewer tonnes. An intensity target commits to fewer tonnes per unit of output or revenue, which is compatible with total emissions rising as the business grows.
Reductions or offsets? A reduction changes what the organisation emits. An offset is a claim about emissions avoided or removed elsewhere, and its quality ranges from rigorous to fictional. A target met largely through offsets is a different commitment.
Additional generation or unbundled certificates? Renewable certificates detached from any specific project move paperwork without adding generating capacity. An agreement financing a wind or solar installation that would not otherwise exist changes the physical outcome. The phrase “100% renewable” covers both.
Which scopes? Scope 1 is direct emissions, scope 2 purchased energy, scope 3 everything else — supply chain, product use, disposal. For most consumer-facing companies scope 3 is the overwhelming majority, so a target covering only scopes 1 and 2 can sound comprehensive while covering little. Check the baseline year too: a reduction from an unusually high starting point can be largely an accounting artefact.
The gap, and closing it
None of this requires settling whether individual climate action matters. It plainly does, in proportion to its measured size, and those sizes are known rather than guessed. What it requires is refusing the substitution: treating a small action as though it discharged the obligation of a large one.
The gap Wynes and Nicholas named was not between people who care and people who do not. It was between what gets recommended and what has been measured.