Carbon avoidance potential: a five-step calculation project

A climate startup can have a strong product, a credible customer, and a market that appears ready to move—and still lose a serious funding conversation because it cannot explain how its solution changes emissions against a realistic baseline.

The hard decision usually comes earlier than founders expect: do you publish a large impact number based on the market you hope to reach, or a smaller number tied to the customers, units, and operating conditions you can defend today? The first number is more exciting. The second is usually more useful.

That is the central tension in a carbon avoidance potential calculation for a climate startup. You are not simply estimating how much carbon your product “saves.” You are comparing two systems: the emissions associated with the incumbent or reference solution, and the emissions associated with your solution. Then you are making the assumptions, timeframe, market volume, and uncertainty visible enough that another person can challenge them.

This is why avoided emissions—sometimes called Scope 4—need discipline. They describe emissions prevented outside the company’s own value chain. They do not cancel out the startup’s Scope 1, 2, or 3 emissions, and they cannot be used on their own to claim carbon neutrality or net zero.

The calculation is not a marketing flourish. Done properly, it becomes a product-validation tool, a commercial planning tool, and a way to find the weak point in a business model before investors or customers find it for you.

Start with the uncomfortable question: is the solution climate-credible?

Before building a spreadsheet, apply the three eligibility gates described in the WBCSD guidance on avoided emissions. These gates are not administrative decoration. They determine whether the impact claim is worth calculating in the first place.

Gate one: climate action credibility

The first question is whether the solution genuinely contributes to climate action rather than producing a secondary benefit that is being dressed up as one.

A climate startup may reduce energy use, lower material waste, extend product life, improve logistics, or replace a higher-emission process. Those effects can be meaningful. But the claim should follow the mechanism, not the other way around.

For example, a software platform that optimizes commercial refrigeration might reduce electricity consumption. The climate effect depends on what the software changes in practice:

• Does it reduce compressor runtime?
• Does it prevent food spoilage and therefore avoid replacement production?
• Does it shift demand to lower-carbon hours?
• Does it simply make the same operation marginally more efficient while increasing total usage elsewhere?

Each effect has a different baseline and a different emissions factor. If a founder combines them into one large number without separating the mechanisms, the model becomes impossible to audit.

Gate two: alignment with the latest climate science

The second gate asks whether the solution is consistent with credible climate pathways, including pathways aligned with limiting warming to 1.5°C.

This is where some apparently “transitional” business models become difficult to defend. Under the WBCSD approach, a solution should not be applied to the exploration, extraction, mining, production, distribution, or sales of fossil fuels. A product that makes an expanding fossil-fuel system more efficient is not automatically a climate solution because it reduces emissions per unit of activity.

That distinction can be uncomfortable for founders. Efficiency is not the same as alignment. A lower-emission version of an activity can still support the continuation or expansion of an activity that climate pathways require the world to phase down.

The practical test is not whether the product has a green adjective in its pitch deck. It is whether the underlying activity moves in a direction compatible with climate science and whether the solution’s contribution is material within that transition.

Gate three: contribution legitimacy

The third gate is about whether the company can legitimately claim the impact.

This matters in crowded systems where several actors contribute to the same emissions reduction. A heat-pump manufacturer, an installer, a financing platform, a building-management system, and a utility may all influence the final result. They cannot each casually claim the full amount as if the others did not exist.

Contribution legitimacy requires a clear account of what the startup actually changes. If your company supplies the control layer but does not manufacture the equipment, operate the building, or control the energy mix, the claim needs to reflect that role.

A credible impact number is not the biggest number your model can produce. It is the number that survives contact with the rest of the system.

These gates also improve your startup strategy. If the solution fails one of them, the answer may not be to abandon the company. It may be to pivot the customer, the use case, or the impact claim. That is a much healthier pivot than building a business around a number that was never defensible.

The baseline is where most impact models quietly break

The reference scenario—or baseline—is the system your solution is being compared against. It sounds straightforward until you try to define it.

A founder will often say, “Our technology replaces diesel,” or “Our platform prevents emissions from conventional construction.” But which diesel equipment? Which operating hours? Which fuel quality? Which building type? Which construction practice? What would the customer actually have purchased if your product did not exist?

The baseline should represent a plausible alternative, not the dirtiest version you can find.

For a carbon avoidance potential calculation, define the reference scenario across at least four dimensions:

• Function: What service does the customer need? Heating, cooling, transport, storage, processing, construction, or something else?
• Timeframe: What period are you assessing, and how might the incumbent change during that period?
• Geography: Which grid, fuel market, regulatory environment, and supply chain apply?
• Operating conditions: What capacity, utilization, lifetime, maintenance pattern, and failure rate are realistic?

A baseline built from a 15-year-old asset may inflate the apparent impact if the real customer alternative is a newer, more efficient model. On the other hand, using an idealized future incumbent can make a useful product look ineffective. Both errors are common because each can be made to support a preferred story.

Build the baseline around the customer’s actual decision

The most useful baseline is often not “the average industry product.” It is the option the buyer would choose without you.

That may be:

• A diesel generator leased for a construction site.
• A conventional refrigeration system already installed in a supermarket.
• A landfill disposal route available to a waste operator.
• A standard cement blend specified by a procurement team.
• A grid-powered process that the customer can access today.
• A manual workflow that creates more travel, material waste, or rework.

The customer’s decision matters because impact depends on adoption. If your product is theoretically better than the industry average but does not replace the purchase your customer would actually make, the theoretical difference is not the realized impact.

Include the solution’s full life cycle

The solution emissions need the same seriousness as the baseline. Include the emissions associated with producing, transporting, installing, operating, maintaining, and retiring the product or service where those stages are material.

For software, that may mean looking beyond the founder’s laptop and considering data processing, connected hardware, site visits, and the operational changes enabled by the system. For physical equipment, manufacturing and replacement parts may matter. For a marketplace or financing platform, the effect may be indirect and require a careful link between the transaction and the changed physical activity.

The point is not to punish a startup for having a footprint. Every business has one. The point is to avoid presenting a gross avoided-emissions figure as though the solution creates climate benefit without generating any emissions itself.

The five-step WBCSD methodology

The WBCSD Guidance on Avoided Emissions sets out a five-step approach. It is a useful structure for an early-stage company because it forces the model to develop in the same order that the business develops: define the claim, establish the alternative, compare the systems, and only then scale the result.

1. Identify the timeframe

Start by deciding what period the calculation covers.

A project may assess:

• One customer deployment over one year.
• The expected operating life of a product.
• A planned commercial period based on sales forecasts.
• A long-term market opportunity extending toward 2040 or 2050.

Do not mix these timeframes in one headline number. A lifetime impact estimate and a one-year realized impact are different statements. Both may be useful, but they answer different questions.

At the early validation stage, a short operating period is often more credible. You may not yet know whether the product will operate for its assumed lifetime, whether customers will renew, or whether the baseline will remain stable. A measured first-year result can teach you more than a polished 20-year projection.

2. Define the reference scenario

Next, specify what would happen without the solution.

Write it as a system description rather than a slogan. Include the equipment, fuel or energy source, operating profile, geography, expected lifetime, and relevant behavior. If the alternative is not a single product but a range of plausible options, document the range and explain why you selected the central case.

A practical model should make assumptions visible in separate cells or fields rather than burying them inside one output. At minimum, record:

• Baseline units of activity.
• Emissions per unit of activity.
• Expected utilization.
• Duration of use.
• Replacement or failure assumptions.
• Data source and confidence level.
• Conditions under which the baseline would no longer apply.

This is also where customer interviews become more valuable than another round of brainstorming. Ask what the buyer would do if your startup disappeared tomorrow. Ask what they bought last time, what procurement approved, and what constraint would block a lower-carbon option. The answers often reveal a baseline that is less flattering—but much more real.

3. Assess solution and reference life-cycle emissions

Now compare the emissions generated by the incumbent and by your solution across relevant life-cycle stages.

The comparison may involve direct emissions, purchased energy, materials, transport, maintenance, and end-of-life treatment. You do not need perfect data to begin. You do need a transparent hierarchy of evidence.

Use measured data where it exists. Use supplier information where it is available. Use recognized emissions factors or conservative estimates where you lack primary data. Mark assumptions that are uncertain instead of presenting them with false precision.

For each system, ask the same questions:

1. What is produced or installed?

2. What energy and materials does it consume?

3. How often does it operate?

4. What maintenance or replacement does it require?

5. What happens at the end of its useful life?

The comparison should be made on a functional unit: one kilowatt-hour of cooling, one tonne of material processed, one passenger-kilometer, one kilogram of product delivered, or another unit that reflects the service being provided.

A functional unit prevents a common modeling error: comparing a small piece of your solution with an entire incumbent system, or comparing equipment that delivers different levels of service.

4. Assess avoided emissions

Once the two systems are described, calculate the difference between the reference scenario and the solution scenario.

At a simplified level:

Net unit impact = emissions from the reference scenario − emissions from the solution

If the result is positive, the solution has a potential avoided-emissions effect for that functional unit. If it is negative, the solution produces more emissions than the reference scenario under the assumptions used.

That negative result is not a failed spreadsheet. It may reveal that the product is being used in the wrong geography, at the wrong utilization rate, with the wrong energy source, or against the wrong incumbent. It may also reveal that the product is not yet a climate solution at the current stage of technology.

Keep the distinction between gross and net impact clear. If the baseline emits 100 units and the solution emits 40, the net unit impact is 60. The 60 is the relevant difference—not the 100 associated with the incumbent.

5. Assess avoided emissions at company scale

The final WBCSD step is to scale the unit-level result to the company’s actual or expected activity.

A simple expression is:

Company-level avoided emissions = net unit impact × volume of units sold or deployed

The word “volume” needs definition. It could mean units sold, units operating, projects completed, tonnes processed, or another measurable activity. Choose the metric that corresponds to the physical effect.

This is where climate impact modeling becomes a business model test. If the unit impact is high but the company can deploy only a small number of units because installation is slow or expensive, the commercial plan needs to acknowledge that constraint. If the unit impact is modest but the product can scale across millions of use cases, the opportunity may still be significant.

Do not hide adoption assumptions inside the final number. Report the activity separately:

• Net impact per functional unit.
• Number of units sold or deployed.
• Expected utilization.
• Period covered.
• Company-level result.

That structure lets an investor, customer, or internal team see whether the impact comes from product performance or from an aggressive volume forecast.

Project Frame: separate potential, planned, and realized impact

The five-step methodology gives you the calculation logic. Project Frame adds a useful distinction for how much confidence to place in the result.

It separates greenhouse-gas impact into three categories: potential, planned, and realized.

Potential impact

Potential impact describes a long-term scenario, often looking toward 2040 or 2050, in which the solution reaches its Serviceable Obtainable Market.

This is a strategic sizing exercise. It can help explain why the category matters, what the solution could contribute at scale, and how the market might evolve.

But potential impact is not a forecast of what the startup will achieve. It is a scenario based on the assumption that the product takes significant market share and that the market conditions support that adoption. Presenting it as current or guaranteed impact is one of the fastest ways to undermine trust.

Planned impact

Planned impact covers a nearer commercial horizon—typically seven to ten years—based on a realistic business plan and sales forecast.

This is the number most relevant to fundraising and operating decisions. It should connect to the company’s go-to-market plan:

• How many customers can the sales team reach?
• How long does deployment take?
• What is the manufacturing or installation capacity?
• What funding is needed to reach the stated volume?
• Which customer segment adopts first?
• What evidence supports the conversion and retention assumptions?

A planned impact model is where the climate claim meets the messy trade-offs of building a company. You may discover that the largest-emitting customer segment has the longest procurement cycle, while the easiest early adopters produce a smaller unit impact. That is not a reason to massage the model. It is a reason to decide deliberately whether the first market is a wedge or a distraction.

Realized impact

Realized impact is historical and based on real-world data.

For an early-stage startup, that might begin with a handful of customer deployments. You can measure baseline activity, solution performance, utilization, operating conditions, and any rebound or substitution effects. The sample may be small, but the evidence can still be more valuable than a large theoretical market estimate.

Realized data also reveals operational problems that the original model missed:

• Customers use the equipment less often than expected.
• Operators bypass the optimization software.
• Maintenance reduces performance.
• The solution replaces only part of the incumbent process.
• A product is installed in a higher-carbon or lower-carbon context than planned.
• The customer keeps the old system running as backup.

These details are not annoyances around the model. They are the model becoming real.

A strong climate startup may present all three, but it should label them plainly. The story becomes more credible when the reader can see the distance between aspiration, plan, and proof.

Use the model to improve the product, not just the pitch

Founders sometimes treat avoided emissions as an impact-reporting exercise that begins after product-market fit. That misses its most practical use.

A carbon impact model can help you decide which version of the product to build.

Suppose you are developing a platform that reduces energy use in industrial facilities. You may initially target every industrial customer. The model could show that the product has a meaningful net unit impact only where equipment utilization is high, the grid is relatively carbon-intensive, and operators act on the recommendations. That changes the first customer profile.

Or suppose you are building a lower-carbon material. The model may show that manufacturing emissions are lower than the incumbent, but transport erases much of the benefit when the product travels long distances. The business may need regional production, a different customer segment, or a denser product format.

This is the value of a lean climate startup approach: treat the impact claim as a hypothesis and test the physical assumptions behind it.

A useful early validation cycle looks like this:

1. Choose one narrow use case. Avoid modeling an entire sector before you have a clearly defined customer and function.

2. Document the customer’s current alternative. Use operational data, invoices, equipment specifications, or observed behavior where possible.

3. Measure the solution under real conditions. Record utilization and performance, not only nameplate capacity.

4. Calculate the net unit impact. Keep baseline and solution emissions separate.

5. Test sensitivity. Change the assumptions that matter most: utilization, lifetime, energy source, adoption rate, or replacement frequency.

6. Decide what must improve. The output should influence product design, pricing, sales focus, or deployment strategy.

This work can be done with a spreadsheet at the beginning. Specialized tools may become useful later, but no software can repair a vague functional unit or an invented baseline. Even a model built while validating a concept—whether through customer interviews, pilot data, or a climate-tech validation tool—still depends on the founder defining the physical change with care.

Where greenwashing enters the spreadsheet

Greenwashing does not always begin with a deliberately dishonest founder. It often enters through a chain of optimistic, individually plausible assumptions.

The baseline is chosen from an unusually inefficient incumbent. The solution is modeled at perfect utilization. The product’s own manufacturing emissions are excluded because they are inconvenient to estimate. The forecast assumes the entire market is available. Then the resulting number is placed next to a much smaller realized figure without explaining the difference.

Watch for these failure modes:

Counting the same impact more than once

If a manufacturer, installer, software provider, and financier all claim the full avoided emissions associated with one deployment, the total narrative becomes inflated. Define the startup’s role and explain whether the claim is shared, attributed, or limited to a specific contribution.

Treating potential impact as delivered impact

A 2050 scenario is not evidence that emissions have already been avoided. Label potential, planned, and realized results separately in investor materials, customer proposals, and public reporting.

Ignoring rebound effects

A product that makes an activity cheaper or more efficient may increase usage. More efficient cooling can lead to longer operating hours. Lower-cost transport can increase demand. If the solution changes behavior, include that possibility in the analysis rather than assuming every efficiency gain becomes an absolute reduction.

Choosing a baseline that will not remain valid

An incumbent may decarbonize, improve efficiency, or be replaced by regulation before your solution reaches scale. The baseline should be reviewed as market conditions change. A business model that depends on a permanently high-emission alternative is fragile by design.

Using avoided emissions to cancel the company’s own footprint

Avoided emissions occur outside the startup’s value chain and must be reported separately from Scope 1, 2, and 3 emissions. They cannot be subtracted from those emissions to claim carbon neutrality or net zero.

That boundary is not a minor technicality. A startup may have a product with substantial avoidance potential and still need to reduce its own operational and supply-chain emissions. Both statements can be true.

Scope 4 can show what your product changes in the world. It does not erase what your company puts into the world.

Make uncertainty part of the result

Early-stage founders often worry that showing uncertainty will make the company look weak. In practice, hiding uncertainty makes the model look immature.

Use ranges where the assumptions genuinely vary. Show a conservative case, a central case, and an upside case if the differences are driven by identifiable conditions. Explain which inputs are measured and which are estimated.

The most useful sensitivity analysis is not a random collection of scenarios. It identifies the variables that can change the investment or product decision.

For instance, if the result changes dramatically when utilization moves from 40% to 70%, utilization is a commercial and operational priority. If the result barely changes across a range of grid-emissions factors, stop spending weeks refining that input. If the product is climate-positive only when it lasts 15 years, product durability becomes central to both engineering and customer success.

A practical internal table might track:

This turns measuring startup carbon avoidance into an operating discipline rather than a one-time claim.

The founder’s real decision

The most important output of the calculation is not a large avoided-emissions figure. It is a sharper answer to three questions:

1. What physical activity changes because this company exists?

2. Compared with what credible alternative?

3. At what scale can the team deliver that change without pretending the messy parts away?

The calculation may show that the original concept is weaker than expected. That can be valuable. You may need to change the target customer, narrow the promise, redesign the product, or find a co-founder who understands deployment, industrial operations, measurement, or life-cycle accounting. Climate entrepreneurship is full of these trade-offs because the product has to work in the physical world, not only in a pitch deck.

The strongest teams do not treat that friction as a distraction from growth. They use it to find the business model that can survive growth.

Start with one functional unit, one customer use case, one defensible baseline, and one timeframe. Calculate the solution’s full life-cycle emissions. Separate potential, planned, and realized impact. Keep avoided emissions outside the company’s own footprint. Then revisit the assumptions whenever real customer data gives you a better answer.

That is the hard-earned lesson: climate impact is not credible because the number is ambitious. It is credible because the path from product behavior to emissions difference is clear, measurable, and honest about what has not been proven yet.