Remove Mutually Exclusive Outcomes From Your Forecast Total
Prepared with AI assistance by Alactic Editorial. Examples identified as hypothetical are illustrative, not customer results.
Hypothetical scenario: A buyer has a standard-package quote, an advanced-package quote, and a replacement proposal. Your pipeline shows three opportunities. The buyer, however, intends to sign only one. Adding all three amounts measures the proposals you have issued, not the revenue the buyer could actually purchase.
To avoid double counting alternative deals in sales forecasts, group records that compete for the same purchase, list the combinations that can actually happen, and calculate the forecast from those outcomes. Keep genuinely additive purchases separate. An alternative should remain visible for selling purposes without becoming another independent contribution to the total.
The distinction matters even in a weighted forecast. Applying a probability to every record does not, by itself, resolve economic overlap. The probabilities must describe compatible possibilities, on a consistent revenue basis and within the same forecast period.
1. Establish what the buyer can purchase together
Start with the buying decision rather than the number of CRM records. Two offers are economically exclusive when accepting one rules out accepting the other for the scope being forecast.
Useful evidence includes a buyer's stated package choice, a single procurement award, a proposal that supersedes an earlier scope, or an agreed implementation plan that supports only one route. Sharing an account name is not sufficient evidence. A customer can buy several unrelated things at once.
Ask the opportunity owners three questions:
- Can both purchases qualify on our chosen forecast basis? Consider the actual scope, not just whether two contracts could technically exist.
- Would one purchase replace, absorb, or cancel the other? If so, identify the amount that would disappear.
- What buyer evidence supports that relationship? Record the confirmation, its date, and who owns the next check.
Alternatives often have different prices, owners, or close dates. Those differences do not make them additive. Conversely, two purchases can share a sponsor or budget discussion without being exclusive.
If the relationship is uncertain, make that uncertainty visible. Do not silently treat every record as independent, and do not delete a legitimate purchase merely because it resembles another one.
2. Set one valuation basis before comparing outcomes
A valid outcome model still produces a misleading total if its amounts measure different things. Choose the forecast period and value basis first. A hypothetical team might use signed first-year contract value for contracts expected to close this quarter. Apply the chosen definition to every route.
Do not compare a full replacement contract with only the incremental uplift on an alternative package unless that is an intentional, documented measurement rule. Similarly, avoid mixing recurring value, total multi-year value, and one-time charges in a single total without a clear reconciliation.
Replacement purchases need particular care. Replacing a buyer's system from another supplier does not automatically create overlap in your forecast. Replacing your own open renewal or earlier proposal might. The relevant question is which amounts in your forecast would coexist after the buyer's decision.
Where an existing renewal is already counted, choose a consistent treatment: forecast the complete replacement outcome while removing the displaced renewal contribution, or forecast a clearly defined incremental amount against a separately maintained baseline. Do not count the complete replacement and the displaced baseline together.
3. Filled hypothetical outcome table: two packages and a replacement
Fictional worked example: A seller is forecasting one buyer's customer-service project. All amounts below are hypothetical US dollars, measured as signed first-year contract value in the same quarter. These are not Alactic customer results or probability benchmarks.
The buyer is considering two packages that extend its existing setup. A third proposal would replace that setup instead. The buyer has confirmed that it will choose at most one route for this project. The earlier packages remain available if the replacement proposal is declined; they have not been withdrawn.
The replacement proposal includes the same project scope covered by the packages. It is not a separate purchase that can be added to either package. No existing renewal from this buyer is included elsewhere in this example's forecast.
Reuse the following filled artifact by replacing the candidate records, buyer evidence, valid combinations, amounts, and probabilities. Keep the no-purchase outcome and reconciliation checks.
HYPOTHETICAL OUTCOME GROUP: PROJECT-17 Period: Current quarter Basis: Signed first-year contract value, USD Decision: Choose one project route, or make no purchase this quarter Owner: Account owner Evidence: Buyer confirms one route only Next check: Reconfirm at next scope review Candidate records A | Standard extension package | $40,000 B | Advanced extension package | $60,000 R | Replacement purchase | $50,000 Filled outcome table Outcome A B R Value Probability EV contribution A only Yes No No $40,000 30% $12,000 B only No Yes No $60,000 25% $15,000 R only No No Yes $50,000 20% $10,000 No purchase No No No $0 25% $0 TOTAL 100% $37,000 Combination validity None Valid: no contract signed this quarter A only Valid B only Valid R only Valid A + B Invalid: alternative packages for the same scope A + R Invalid: replacement includes the package scope B + R Invalid: replacement includes the package scope A + B + R Invalid: combines competing routes Reconciliation Raw candidate-record sum: $150,000 Maximum feasible group value: $60,000 Expected group value: $37,000 Probability of any purchase: 75% Value conditional on a purchase: $49,333, rounded
The expected value is 30% × $40,000 + 25% × $60,000 + 20% × $50,000 = $37,000. The no-purchase outcome contributes zero but remains essential: it represents the possibility that nothing closes in the period.
The conditional value is $37,000 divided by 75%, or approximately $49,333. It answers a different question: what is the average value if a purchase happens? It is not the unconditional forecast contribution.
The $150,000 raw total is useful only as a sum of candidate proposal values. It is not achievable project revenue under the stated constraints. The $60,000 figure is the maximum feasible value, not a prediction. The $37,000 figure is an expected value based on the stated hypothetical probabilities, not a promised result.
4. Make probabilities describe the decision, not isolated records
For a group in which only one outcome can occur, the probabilities of its purchase outcomes cannot total more than 100%. Adding a no-purchase outcome makes the full set total exactly 100%.
Generic sales-stage weights may fail this test. Multiple alternative proposals can each be far advanced in the selling process without each representing an independent chance of revenue. Proposal maturity and the probability of winning a particular route are not the same measurement.
One workable approach is to estimate two things separately:
- Purchase probability: How likely is any qualifying purchase in this period?
- Route share: Conditional on purchasing, how likely is each mutually exclusive route?
Multiply the purchase probability by each route share to obtain its unconditional outcome probability. Route shares must total 100% among the available purchase routes. Keep the remaining probability as no purchase in the forecast period.
Importantly, summing weighted records is not inherently wrong. If the record probabilities are already coherent, unconditional probabilities for these exclusive outcomes, their weighted sum gives the same $37,000 expected value in the hypothetical example. The defect is treating overlapping records as independently achievable revenue, not the mere presence of multiple records.
Nor should every reduction in a weighted forecast be attributed to removing duplication. If you change both the outcome structure and the probability judgments, document those as separate changes.
5. Distinguish alternatives from genuinely additive purchases
A blanket rule such as taking only the largest opportunity per account solves one error by creating another. It can suppress purchases the buyer really can make together.
Hypothetical contrast: Suppose the same buyer can purchase an independent training engagement worth $8,000 alongside any project route. If the engagement has separate scope and is not included in the package or replacement price, its value may be additive. Confirm that it can genuinely coexist rather than assuming this from its separate record.
Conditional purchases need more care. In this hypothetical example, if training is available only when the advanced package is selected, training alone is not a valid outcome. Model the advanced package with and without training, or calculate the training contribution using a probability that already accounts for the advanced-package dependency. Do not apply a standalone training close probability that ignores its prerequisite.
Partly overlapping scope also calls for more than a simple exclusive flag. Separate the common scope from independently purchasable additions, or enumerate the valid bundles. Otherwise, treating everything as additive overstates revenue, while treating everything as exclusive understates it.
For genuinely additive purchases, expected values can be added even when their likelihoods are related, provided each contribution uses a coherent unconditional probability. A combined best-case scenario, however, still needs a feasibility check. Shared budget or capacity constraints may prevent all purchases from happening together.
6. Preserve selling detail without counting it twice
You do not need to erase alternative proposals to produce a disciplined forecast. Keep the commercial detail needed to manage negotiations, but distinguish proposal visibility from forecast contribution.
A recommended operating design uses a small set of shared fields or an equivalent maintained worksheet:
- Outcome-group identifier: links competing routes to the same buying decision.
- Relationship: alternative, replacement, additive, conditional, or unresolved.
- Value basis and period: prevents incompatible amounts from being combined.
- Validity rule: states which combinations are allowed and why.
- Evidence and review date: supports the relationship judgment.
- Forecast treatment: identifies whether the total uses a group contribution or individual record contributions.
Choose one aggregation path. Either calculate the group total from its outcomes, or sum properly weighted member contributions. Do not add the group summary on top of its members.
If your reporting cannot represent outcome groups, maintain a controlled reconciliation outside the raw pipeline report. Show the raw candidate amount, the feasible maximum, and the expected contribution separately. Make exclusions traceable rather than silently changing the face value of a live quote.
These are recommended operating practices, not assertions that a particular CRM provides these controls. Check your workspace and current plan before designing a workflow around any specific feature.
7. Recover when the buying decision changes
The model should change when the buyer's choices change. Reassess it whenever scope, funding, timing, or the availability of a route changes.
Use an explicit update sequence:
- Recheck scope. Establish whether the routes still compete for the same purchase.
- Update valid combinations. Remove impossible outcomes and add newly feasible ones.
- Reassess probabilities. Do not automatically transfer a withdrawn route's probability to the largest remaining deal. Some of it may belong to no purchase.
- Reconcile the total. Explain changes in scope, price, probability, and timing separately.
- Retain the prior snapshot. Preserve enough history to explain why the forecast moved.
If an earlier quote is formally superseded and cannot be accepted, it is no longer a live outcome. Retain it for commercial history, but remove its current forecast contribution according to your reporting convention.
If the relationship cannot be resolved before a forecast meeting, label it unresolved and show the competing interpretations. Do not present an additive total as established fact merely because it is easier to produce.
8. Test the next forecast with a small reconciliation
For the next forecast review, select a manageable set of accounts with multiple open offers for similar scope. For each, complete the same outcome table used above before changing the total.
Measure three observable results: the number of groups with confirmed combination rules, the amount removed from an infeasible additive scenario, and the number of unresolved relationships with a named owner and review date. Keep any expected-value change separate from the reduction in the raw proposal sum.
Then run four checks: all outcome probabilities sum to 100%; no invalid combination contributes revenue; all amounts share a defined basis; and no group summary is counted alongside its members.
The practical target is a forecast in which each contribution corresponds to an economically possible purchase. Alternative offers can remain visible and useful without pretending that the buyer will purchase them all.
Put the next conversation to work.
Explore Alactic Communications for WhatsApp conversations, campaigns and AI-assisted workflows. Choose the product and plan that fit your team.
Explore Communications