After this lesson you will be able to
- Collect the five or six inputs that drive a Salesforce business case directly from the customer.
- Convert revenue and time savings into gross margin and real cost, the way finance will.
- Run a simple sensitivity check so the case survives skeptical questions.
- Produce a one-page summary that a champion can take into a budget meeting.
Start from the customer's numbers, not yours
The fastest way to lose a business case is to fill it with numbers the customer has never seen. Industry averages and vendor benchmarks invite the question, "Is that true for us?" and nobody in the room can answer it.
Build the case in a working session with your champion, ideally with someone from finance or operations who knows the numbers. Ask for the inputs, write them down in front of them, and label any number that is an estimate. When the customer supplies the inputs, the case stops being your claim and becomes their analysis.
If they cannot give you a number, ask for a range. "Somewhere between two and five hours a week per rep" is a perfectly good input. Use the low end in the base case.
Pick two or three value drivers and ignore the rest
Salesforce can affect dozens of metrics. A business case that lists all of them looks padded and is impossible to defend. Choose the two or three drivers tied to the business problem you found in discovery, where the customer can name an owner and a current baseline.
Each driver needs a baseline today, a realistic change, and a person who will be accountable for that change. If nobody will own a benefit, leave it out. It will not happen, and the CFO knows it.
- Time recovered: hours per week per person on manual work, multiplied by loaded cost, and only counted if the time is redeployed
- Win rate or conversion: change in close rate on a defined pipeline, converted to margin
- Retention: reduction in customer churn, valued at the margin on retained revenue
- Service cost: cases handled per agent, or cases resolved without an agent
- Avoided cost: tools retired, contractor spend reduced, licenses consolidated
- Cash timing: faster quote-to-cash or billing accuracy that shortens days to collect
Count margin, not revenue
This is one of the most frequent errors in Salesforce business cases. If better pipeline management adds a million dollars in revenue, the business does not gain a million dollars. It gains the gross margin on that revenue, after cost of goods or delivery.
Ask the customer for their gross margin percentage, or a range. Multiply every revenue benefit by it. For example, a manufacturer with a 30 percent gross margin that expects $500,000 in additional annual revenue should show a $150,000 annual benefit, not $500,000. The smaller number is the one the CFO will believe.
Treat time savings with the same discipline. Hours saved are only worth money if they turn into more selling, more service capacity or avoided hiring. Say which one, and say it plainly.
Put every cost on the page
The return is only half the case. On the cost side, include subscription fees over the term, implementation services, data migration, integrations, AgentExchange (formerly AppExchange) apps, training, the customer's own internal time and ongoing administration. If Agentforce is in scope, include usage-based consumption such as Flex Credits as a separate line with its assumptions.
Ranges are fine. Hidden costs are not. If services are uncertain, show a low and high estimate and explain what drives the difference. Our pricing guide at /resources/salesforce-consulting-pricing-guide/ helps customers understand typical services structures. When you need a grounded services estimate for a specific deal, a delivery partner can scope it; Abstrakt does this through the deal desk at /salesforce-account-executives/deal-desk/.
Stress-test it with sensitivity, before the CFO does
A single ROI number invites a single question: what if you're wrong? Answer it in advance. Run the case three ways: a base case with the customer's own conservative inputs, a low case where adoption is slower and benefits are smaller, and a high case only if the customer insists on one.
The most useful sensitivity test is simple. Cut every benefit in half and push the start date back by a quarter. If the project still pays back within the contract term, say so. That one sentence carries more weight than any precise ROI percentage.
Also find the break-even point. "This pays back if reps recover at least 90 minutes a week" gives the customer a threshold they can judge from their own experience.
Fit it on one page and let the customer own it
The output is a one-page summary, not a spreadsheet. Lead with the problem, then the investment over the term, the annual benefit in margin terms, payback period, the low-case result, the key assumptions and the named owners. Put the detailed model behind it for anyone who asks.
Our ROI calculator at /salesforce-account-executives/roi-calculator/ is built for this. Pick a scenario, replace the placeholder values with the customer's numbers, and it produces a three-year view of benefit, cost, ROI and payback with every formula visible. It counts margin rather than revenue by design, and it prints cleanly for a budget meeting.
Hand the final version to your champion with their company name on it and your name off it. A business case presented by an internal leader is a plan. One presented by a vendor is a proposal.
Talk tracks
Proposing the business case session
“I'd rather not hand you a vendor ROI slide. Can we spend 45 minutes building the numbers together with your figures, so you can defend every line?”
It positions the case as the customer's analysis and gets you access to real data.
When the customer does not know a number
“No problem. Give me a range you'd be comfortable defending, and we'll use the low end. We can tighten it later.”
Ranges keep the work moving and conservative inputs build credibility.
When your champion wants to count revenue
“Your CFO will ask what that revenue is worth after delivery costs. What's your gross margin, roughly? Let's show that number so nobody can poke a hole in it.”
It protects your champion from the most common finance objection.
When someone asks "what if this doesn't happen?"
“We tested that. If every benefit comes in at half and we start a quarter late, it still pays back in about this many months. Here's the break-even point.”
Sensitivity turns doubt into a threshold the customer can judge.
When a benefit has no owner
“I'd suggest we leave this one out. If nobody here would sign up to deliver it, it shouldn't be in the case.”
Removing weak benefits makes the remaining ones more believable.
Handing over the final summary
“This is yours now. Put your logo on it and change anything that doesn't sound like you. I'm happy to join the review if it helps, but you should present it.”
Internal ownership is what gets a business case approved.
Business case build sheet
Use it on a live deal. Tick items off as you go; nothing you tick leaves this page.
Common mistakes
Using vendor benchmarks as inputs
Generic benchmarks invite "is that true for us?" and nobody can answer. Use the customer's numbers, even if they are rough.
Counting revenue as return
Revenue overstates the benefit by the cost of delivering it. Always convert to gross margin.
Counting saved time that is not redeployed
Hours saved only matter if they turn into selling, capacity or avoided hiring. Otherwise finance will discount them to zero.
Leaving out internal and services costs
A case that shows only license cost will be rebuilt by finance with bigger numbers. Show the full cost first.
Presenting it yourself
A vendor-presented ROI reads as a sales tool. Make sure the champion presents it as their plan.
How you will know it is working
- The customer edits the case and adds numbers you did not ask for.
- Finance challenges individual assumptions instead of rejecting the whole model.
- Your champion presents the case without you in the room.
- The low-case result still clears the customer's payback threshold.
Questions
How long should building a business case take?
One working session of about an hour to collect inputs, plus a short review once the summary is drafted. If it takes much longer, you probably have too many value drivers. Narrow to the two or three tied to the core problem.
What payback period do customers expect?
It varies by company and by finance leader, so ask. Many will tell you their threshold if you ask directly. Build the case to show whether the low case meets it.
Should I include soft benefits like better visibility?
Mention them in a short qualitative line, but do not put a dollar value on them unless the customer can name a specific decision or cost they affect. Unquantified soft benefits support a case. Inflated ones undermine it.
How do I build a case for Agentforce specifically?
The same rules apply, with one addition: consumption costs depend on volume, so model them from the customer's actual case or conversation volumes. Our guide at /resources/agentforce-roi-business-case/ covers the Agentforce-specific inputs.