Plan on first measurable gains within 90 days if you scope tightly, break-even somewhere around months 12 to 18, and the bulk of the return in years 2 and 3. That’s our planning range, and it carries no guarantee. A Salesforce growth and efficiency roadmap that ranks priorities before the build gives that range a starting point.
Three Salesforce-commissioned Forrester studies report payback in under 6 months, but each models a composite company, and the manufacturing one started Year 1 at 25% adoption with implementation taking 7 to 24 months. In our worked example, a 40% ROI on paper turns into break-even in month 14 once the adoption ramp counts.
Four things drag payback: adoption, inherited data, late integration and a slow first release. Check logins at 30 days, pipeline logged in Salesforce at 90 and hours saved against plan at 180, and put your own break-even month in front of leadership on day one.
TL;DR
- Forrester’s three studies report 396% ROI for Agentforce in customer service, 354% for manufacturing and 122% for Industry Clouds.
- The Industry Clouds composite booked no benefit in its initial period, then grew benefits about 80% between Year 1 and Year 3.
- The Agentforce composite rested on case deflection rising from 25% to 35% and a 50% cut in handling time on transferred cases.
- In the worked example, half-speed adoption after month 6 pushes break-even from month 14 to month 35.
- A baseline set before kickoff is what makes the 30, 90 and 180-day numbers provable.
What Do the Published ROI Studies Say?
Salesforce commissioned all three studies, and each one reads payback off a composite company’s three-year cash flow. A composite blends several interviews into one company, and your calendar will differ.
Agentforce for Customer Service (November 2025). The composite has 50 service reps and 500,000 cases in Year 1, built from 6 interviews, and it reports 396% ROI with payback under 6 months. Interviewed companies implemented in 2 weeks to 6 months, and the composite’s implementation effort was 550 hours.
Salesforce for Manufacturing (July 2024). The composite has $2.3 billion in revenue and 4,300 employees, built from 5 interviews, and it reports 354% ROI with payback under 6 months. Adoption runs 25%, 50% and 90% over 3 years, and interviewed companies took 7 to 24 months to implement.
Industry Clouds (August 2026). The composite has $3 billion in revenue and 10,000 employees, built from 11 interviews, and it reports 122% ROI with payback under 6 months. Deployment is phased, starting with high-impact business units, and the initial implementation ran 6 months.
Check the interview counts before you cite any headline: 5 interviews is a thin base for a 354% ROI, and a composite can look nothing like your business. The Agentforce composite had a light implementation and benefits from Year 1, which is why payback came fast, and a project with a long build and a slow rollout won’t repeat that curve. The Industry Clouds composite shows the other shape: its first period returned nothing, and benefits climbed after that, so the open question is how long your quiet stretch lasts.
A Month-by-Month ROI Example
Here’s an illustrative rollout, with spend indexed to 100 so the shape of the curve matters more than the currency. Swap in your own numbers.
- Year 1 spend: 100, covering the platform, the build, internal time and support
- Benefit at full adoption: 140.4 a year, from 5,520 hours saved (60 sellers × 2 hours a week × 46 weeks) plus retired tools
- Monthly benefit at full adoption: about 11.7
- Ongoing spend after Year 1: about 3.7 a month
Finance will count the retired tools as cash and the 5,520 hours as a maybe, until those hours turn into more selling or less overtime. Count the hours, but keep them on a separate line.
Simple ROI is (140.4 minus 100) divided by 100, or 40.4%. That’s the number a proposal shows.
Now add the ramp. Benefits run at 0% in months 1 to 3, 50% in months 4 to 6 and 100% from month 7, which yields 87.8 of Year 1 benefit against 100 of spend, so cumulative cash sits at minus 12.2 at month 12.
From month 13 the ongoing spend of 3.7 a month meets 11.7 of monthly benefit (a simplification, since admin and support effort continues in practice), so the line crosses zero in month 14. The same project that returns 40% on paper is still 12% of its spend short after a year.
Change one assumption and the answer moves: if adoption stalls at 50% after month 6, Year 1 benefit drops to 52.7 and break-even slips to month 35. If the first release lands a month early, break-even comes in month 13, and if go-live slips 2 months it moves to month 17. That sensitivity tells you which assumption to protect.
What Slows Salesforce ROI Down
Adoption. The manufacturing composite assumed only 25% adoption in its first year, and every benefit in your model depends on people using the system. The Industry Clouds composite grew from 20% of its employees as users to 50% in Year 1, 75% in Year 2 and 100% by Year 3.
Data you inherit. Duplicates and gaps from the old system surface the first time someone builds a report, and cleanup eats the weeks your benefit was meant to start.
Integration found late. An ERP connector scoped after go-live pushes the benefit start date back by however long that connector takes to build.
A slow first release. Benefit stays at zero until something goes live: in the example, months 1 to 3 return nothing, and a 2-month slip in go-live adds 3 months to break-even.
All four come from how the rollout is run, which is why two companies on the same platform can sit months apart on payback. Fix the ones you control in the first 90 days, because every later number inherits their gaps.
What to Track at 30, 90 and 180 Days
Early numbers tell you whether the later ones will arrive. Each window below answers one question:
- 30 days: login rate and required-field completion show whether people use the system at all
- 90 days: share of pipeline logged in Salesforce and duplicate rate show whether the data can be trusted
- 180 days: forecast accuracy, hours saved against plan and retired tools show whether the model’s benefits are arriving
If a 30-day number comes in low, fix it before the 90-day review. Set a Salesforce ROI baseline before kickoff, because you can’t prove 2 hours saved a week without a number for last quarter, and then pick the first release by what you can measure against that baseline.
How Do You Shorten the Wait?
1. Release narrow. One workflow most of the team uses beats five that half of them route around.
2. Roll out by team. Phased deployment is how the Industry Clouds composite ran, and each wave gives you a real adoption number to plan from.
3. Name one owner for adoption. Someone has to chase the logins and report them every week without being asked.
4. Re-forecast at day 90. Replace the planned adoption curve with the real one and update the break-even month in your reporting when it moves.
Put month 14, or whatever your own model says, in front of leadership on day one, so a late break-even never arrives as news. Then report the 30-day login rate in week 5, while there’s still time to change something.