Most business cases are good at describing the future with the project: the extra sales, the saved hours, the new customers, the lower costs. Far fewer are as careful about the future without it. That imbalance matters, because the value of a decision is not what happens after it. It is the difference between what happens with it and what would have happened otherwise.
Some of the benefit claimed for an investment would have arrived anyway. Some is moved from another part of the business rather than created. Some leaks to suppliers, platforms or customers. And the money, people and attention committed to the project are no longer available for anything else. The missing question in many business cases is the counterfactual: what would plausibly happen if we did not do this, or did something different instead?
This article explains how to build a realistic base case, how to separate genuine new value from benefits that would have happened anyway or were moved from elsewhere, how to count costs that never appear on an invoice, and how to measure after implementation whether the forecast was right.
Sequence is not cause
When results improve after an investment, it is tempting to credit the investment. But sequence does not establish cause. Sales may have risen because the market grew. Costs may have fallen because input prices dropped. A new channel may show strong sales because existing customers moved to it.
Research by Dennis Coates and Brad Humphreys on publicly funded sports stadiums in the United States illustrates the problem at a large scale. Optimistic forecasts made before construction often predicted large gains in local income and jobs. Studies looking back afterwards frequently found little or no overall gain. One reason was substitution: people spent money at the stadium instead of at restaurants, cinemas and other local venues. Another was leakage: much of the money flowed to players, owners and others outside the local area. The findings belong to their setting. The lesson about evidence applies to any business case.
Gross benefit is not incremental benefit
A new online store that takes $400,000 of orders has not necessarily created $400,000 of new sales. Some of those customers would have bought in the shop. A self-service portal that handles thousands of enquiries may simply have moved them from the phone. A new product that sells well may have taken sales from an existing one.
The credible benefit is the difference between two realistic futures: with the investment and without it. Three questions help:
- Additionality: what changes because of this investment that would not have changed otherwise?
- Attribution: how much of that change can reasonably be credited to this investment rather than to other factors?
- Opportunity cost: what value is given up by using money, people and attention here rather than elsewhere?
The base case is not a frozen present
Many business cases compare the project with a “do nothing” scenario in which everything stays exactly as it is today. That is rarely realistic. Without the investment:
- equipment ages and maintenance costs rise;
- competitors move;
- customer expectations change;
- costs and prices change;
- the business would probably do something, even if smaller.
A realistic base case describes what the business would actually do and how conditions would evolve without the project. Sometimes that makes the project look better than a frozen comparison, because the base case is deteriorating. Sometimes it makes the project look worse, because a cheaper “do minimum” option would achieve much of the benefit. Either way, it gives a more honest answer.
Substitution and cannibalisation
Substitution occurs when the benefit of the project replaces activity that would have happened anyway, inside or outside the business. In a business, it often appears as cannibalisation: a new product, channel or location taking sales from an existing one.
Common examples:
- a new website taking orders that would have come by phone or in person;
- a second location drawing customers from the first;
- a premium product range drawing buyers away from the standard range;
- a central purchasing saving offset by higher stock levels or longer lead times elsewhere.
None of these is necessarily bad. A business may choose to cannibalise its own sales before a competitor does. The point is to count only the net change. The who wins, who pays article discusses how moved costs and benefits can be mistaken for new value.
Leakage: who captures the value
Some of the value created by a project may flow to others: a platform charging fees on every sale, a payment provider, a supplier who raises prices once volume grows, or customers who receive a benefit without paying more for it. That is not necessarily a reason to reject the project, but it should be counted. Ask who actually receives each benefit.
Costs without an invoice
Opportunity cost is real even when no invoice exists. A team working on one project cannot work on another. The owner’s time spent on a new venture is time not spent with existing customers. Cash committed to one investment cannot fund another. Business cases often ignore these costs because they do not appear in the project budget. Name the most valuable thing the business would give up, and include an estimate of it.
Forecasts are hypotheses
A business case forecast is a hypothesis about what will happen. It is not evidence that it did. Many businesses review business cases carefully before approval and never check them afterwards. Sponsors move on, data becomes hard to find and benefits are reported selectively. Without that check, optimism never gets corrected.
Three habits help:
- Set the baseline before implementation. Record current sales, costs, times or volumes before the change. A baseline built after results are known is easy to bend.
- Use a comparison where practical: a pilot group compared with a similar group that did not change, a phased rollout, or a comparison with the trend before the change.
- Keep a forecast record. Over time, compare what business cases predicted with what happened. Which benefits are usually overstated? Where does adoption take longer than expected? Which costs are usually forgotten?
Watch for double counting across projects
When several initiatives run at once, they often claim the same benefit. A new scheduling system, a training program and an extra vehicle might each claim credit for the same reduction in late deliveries. A pricing review and a new product range might both claim the same margin improvement. Added together, the claims exceed what actually changed. Before approving several related projects, list the benefits each one claims and check where they overlap. Where two projects depend on the same change, count the benefit once and decide which project owns it.
Test how much the answer depends on each adjustment
Counterfactual estimates are rarely precise. Substitution, leakage and the base case trend are all judgements. Rather than pretending to precision, test how the result changes if each estimate is higher or lower. If the project remains worthwhile across a reasonable range, the decision is robust. If it flips from worthwhile to loss-making when one estimate moves a little, that estimate deserves more evidence before the business commits, perhaps through a small trial.
A counterfactual ledger
For any significant business case, work through eight items:
| Item | Question |
|---|---|
| Base case | What would realistically happen without this investment? |
| Alternative | What other credible action could address the same need? |
| Incremental effect | Which benefits and costs change relative to the base case? |
| Substitution | What existing sales, activity or spending would move rather than grow? |
| Leakage | Which benefits flow to others, such as platforms, suppliers or customers? |
| Displacement | What work, investment or attention is crowded out? |
| Evidence | Which claims rest on data, pilots or contracts, and which are assumptions? |
| After-check | What data will show whether the forecast was right? |
A worked example
This is an illustration. A homewares retailer with two shops plans an online store. The business case forecasts $400,000 of online sales in the first year. At a 45% gross margin, that is $180,000 of gross profit, against $90,000 a year for the platform, photography, packing and a part-time staff member. The case claims a net benefit of $90,000.
The owner works through the counterfactual ledger:
- Substitution: from conversations with customers and a similar retailer’s experience, the owner estimates that about a quarter of online sales, $100,000, would come from existing shop customers who would otherwise have bought in store. The margin on those sales, $45,000, would have been earned anyway.
- Leakage: payment and platform fees of about 3% of online sales take $12,000.
- Realistic base case: without an online store, the owner expects to lose some customers to competitors who sell online, roughly $60,000 of shop sales a year. The online store should retain most of them, protecting about $27,000 of margin that a frozen “do nothing” comparison would have missed.
- Displacement: the owner expects to spend about six hours a week on the online store, time currently spent on wholesale accounts. The estimated margin given up is about $20,000.
| Item | Amount |
|---|---|
| Gross profit on forecast online sales | $180,000 |
| Less margin on sales that would have been made in the shops | −$45,000 |
| Less payment and platform fees | −$12,000 |
| Plus margin protected from customers who would otherwise leave | +$27,000 |
| Less running costs | −$90,000 |
| Less owner’s time diverted from wholesale | −$20,000 |
| Credible net benefit | $40,000 |
The project is still worthwhile, but its credible benefit is less than half the original claim, and much more sensitive to the substitution estimate. The owner proceeds with a measurement plan: online orders are matched against the shop customer list to see how many are from existing customers, shop sales are tracked against their previous trend, and the results are reviewed after six months. The forecast becomes something the business can test, rather than a number that is quietly forgotten.
How this applies to a small Australian business
Small businesses often make investment decisions quickly, with simple forecasts. A few extra questions make those forecasts much more reliable:
- Describe the base case realistically, including what would change without the project.
- Ask how much of the benefit is genuinely new.
- Estimate substitution and cannibalisation.
- Count fees and other leakage.
- Name the opportunity cost, especially the owner’s time.
- Record baselines before implementation.
- Compare forecasts with results and keep a simple record.
- Discuss significant investments with your accountant, including tax and cash flow effects.
The articles on cost-benefit analysis for business decisions and when the project succeeds and the strategy fails cover related methods.
Signals worth watching
- Business cases with detailed benefits but a vague base case.
- New channels claiming all the revenue that flows through them.
- “Do nothing” scenarios that assume nothing changes for years.
- The same benefit counted in more than one project.
- No plan to measure benefits after implementation.
- Reviews that confirm the project was delivered but not that it caused the expected change.
Common mistakes
- Comparing against a frozen present.
- Counting gross activity as new value.
- Ignoring cannibalisation.
- Forgetting fees and leakage.
- Leaving out opportunity cost because there is no invoice.
- Never checking whether the forecast came true.
Frequently asked questions
How do we estimate substitution without data? Use whatever evidence is available: customer conversations, the experience of similar businesses, a small trial or a sensible range. State the assumption clearly and test how much the result depends on it.
Is a realistic base case always worse than doing nothing? No. Sometimes the base case is deteriorating, which makes the investment look better than a frozen comparison. The aim is accuracy, not pessimism.
What is a good comparison for a small business? Often the trend before the change, or a comparison between a group that changed and one that did not, such as one shop versus another, or customers who used a new service versus those who did not.
How long after implementation should we check? Long enough for the benefit to appear, often three to twelve months. Set the review date when the decision is made.
What if the after-check shows the forecast was wrong? That is valuable. Adjust the project if needed, and use what you learned to improve the next forecast.
Does this apply to cost-saving projects too? Yes. A saving is also a difference between two futures. Ask whether costs would have fallen anyway, for example because a contract was due for renegotiation, whether the saving moves cost to another area, and whether the hours saved will actually reduce spending or simply be absorbed.
Questions to ask
- What would actually happen if we did not make this investment?
- Which benefits are genuinely new rather than moved from elsewhere?
- Who captures part of the value we expect?
- What are we giving up to do this?
- What baseline will we record before we start?
- How will we know whether the forecast was right?
Bringing it together
The value of an investment is the difference it makes, not everything that happens after it. Describe a realistic base case, separate new value from benefits that would have happened anyway or were moved from elsewhere, count leakage and opportunity cost, and record a baseline so you can check the result afterwards. The strongest business case is not the one with the biggest numbers. It is the one that most credibly explains what the investment will change.
Source: KEVOS notes, drawing on D. Coates and B. R. Humphreys’ research on professional sports facilities and urban economic development, and on G. Davison and colleagues (2020) on stakeholder perspectives in measuring returns. Examples and figures in this article are illustrations. This article is general information, not financial advice.
