Strip 15 to 30 Percent From Your Salesforce Bill Before Renewal
Shelfware runs 15 to 30 percent of paid seats in most Salesforce orgs, and the largest recovery comes from a 90 day login audit and edition right sizing, not from a deeper discount.
Prepared by Redress Compliance · June 2026 · Representative Salesforce estate scenario (benchmark scenario, not a quote)
Executive Summary
Salesforce 2026 list rates set the stage for the waste. Sales Cloud Enterprise is USD 175 and Unlimited USD 350 per user per month, Service Cloud Enterprise is USD 165 and Unlimited USD 330, after the broad 6 percent list increase Salesforce applied in August 2025.
Most of the overspend is not in the rate. It is in seats nobody logs into and in editions that sit a tier above the work the user actually does. Both are invisible on the renewal proposal because the proposal prices what you bought last time, not what you use.
The recovery has a sequence. Reconcile active logins, right size editions, govern sandbox and storage line items, then negotiate the rate. Discount last, not first, because a 30 percent discount on a dormant seat is still money spent.
One contract fact frames the whole exercise. Salesforce subscriptions do not shrink mid term. You can only reduce seat counts at renewal, and only inside the notice window. Miss it and the count rolls forward at the contracted uplift.
Across roughly 30 to 45 Salesforce license reviews we benchmarked between 2024 and 2025, buyers who ran the full sequence recovered 15 to 30 percent of paid seats before renewal. This paper gives the 90 day audit, the edition ladder, the silent line items, the permission set audit risk, and the three year math.
How do you find Salesforce shelfware in 90 days?
You find it in login data, not in the order form. Salesforce records the last login date on every user, and a user who has not logged in for 60 to 90 days is shelfware regardless of what the contract calls them. The 90 day audit turns that signal into a reclaim list before the renewal conversation starts.
Run it as three sprints of one month each. The point of the cadence is to have a defensible, evidence backed seat count in hand before Salesforce sends a proposal, so the negotiation starts from your number.
| Sprint | What you pull | The reclaim signal |
|---|---|---|
| Days 1 to 30 | Last login date, login frequency, and license type for every active user. | Zero or rare logins over 60 to 90 days. Inactive and duplicate accounts. |
| Days 31 to 60 | Edition and feature usage by user. Which Unlimited features each seat actually touches. | High edition users whose work fits a lower edition. Unused add on products. |
| Days 61 to 90 | Sandbox inventory, storage consumption, and permission set license assignments. | Idle full copy sandboxes, storage overage, and over assigned feature entitlements. |
The output is a single reconciled number: the seats and entitlements you will actually renew, each backed by usage evidence. That number, not the prior year count, is the opening position.
Where reclaimable Salesforce spend sits, as a share of total recovery. Shares sum to 100 percent. Benchmark scenario, not a quote.
How do you right size editions across Sales, Service, and Platform?
You map each user to the lowest edition that covers their real work, then hold the line against the upsell. Salesforce sells editions as a ladder, and the gap between rungs is large. Moving the wrong users up the ladder is the second largest source of waste after dormant seats.
These are the 2026 published list rates per user per month. Confirm them against the vendor pages: Sales Cloud pricing, Service Cloud pricing, and the Platform editions.
| Edition | 2026 list, per user month | Fits the user who | Right sizing move |
|---|---|---|---|
| Platform Starter | USD 25 | Uses custom apps and records, not the core CRM objects. | Move light internal users off full CRM seats. |
| Platform Plus | USD 100 | Needs more custom objects and logins than Starter allows. | Home for builders who do not sell or service. |
| Service Cloud Enterprise | USD 165 | Runs cases and the console day to day. | Default for support agents. Resist Unlimited. |
| Sales Cloud Enterprise | USD 175 | Manages pipeline with standard automation. | Default for most sellers. |
| Sales Cloud Unlimited | USD 350 | Genuinely needs the premium support and full feature set. | Reserve for the few who use it. Do not blanket it. |
The lever is the spread. Unlimited is double Enterprise. Every seat parked on Unlimited that only needs Enterprise wastes USD 175 a month, or USD 2,100 a year. A Platform Plus seat that should be Platform Starter wastes USD 900 a year. These numbers add up fast across hundreds of users.
2026 Salesforce list rates per user per month. The Unlimited rung is double Enterprise, which is where edition waste concentrates.
- Map features to roles. List the Unlimited only features and identify the exact users who touch them. Everyone else holds at Enterprise.
- Refuse blanket Unlimited. Reject any proposal that lifts the whole base to Unlimited to cover a minority that needs premium support.
- Push builders to Platform. Internal users who never work leads or cases belong on Platform Starter or Plus, not a full CRM seat.
Why are sandboxes and storage silent line items?
They are silent because they are priced off your spend or your data, not off a user you can see in a login report. A full copy sandbox and a storage overage do not appear on a seat list, so they survive every renewal that only counts users.
The pricing mechanics are the trap. Sandboxes scale with the size of your contract, and data storage is sold in small, expensive blocks well above its marginal cost.
| Line item | How it is priced | Why it inflates | The control |
|---|---|---|---|
| Full Copy Sandbox | Up to 30 percent of net annual spend. | Scales with the estate, not with how often it is used. | Cap as a flat fee. Refresh only at release windows. |
| Partial Copy Sandbox | Up to 20 percent of net annual spend, 5 GB cap. | Quietly duplicated across teams that could share one. | Consolidate. Share a refresh schedule. |
| Data storage | About USD 125 per month per 500 MB above the included tier. | USD 1,500 a year per 500 MB block regardless of cost to serve. | Archive, big objects, external storage. |
| File storage | About USD 5 per month per 1 GB. | Creeps up quietly from email and case attachments. | Offload attachments to an external file store. |
Included storage is modest: most editions ship roughly 10 GB of data storage plus 20 MB per user license, confirmed on the Salesforce storage allocation documentation. A large org blows through that on history alone, and the overage is billed in those USD 125 blocks.
Do permission sets create an audit risk?
Yes, when they assign feature entitlements you are not contracted for. Permission sets and permission set licenses grant access to features such as advanced functionality or add on products. Each assignment is a billable entitlement, and Salesforce can reconcile assigned entitlements against your contracted quantities at renewal.
This is where the standard administrative advice and the licensing reality pull apart. Cleaning up access through permission sets is good hygiene. It is also the surface a vendor reconciliation looks at first.
Where the common advice on permission sets is wrong
The standard reseller and admin advice is to consolidate profiles into permission sets for cleaner administration, and to assign generously so nobody is blocked. We disagree on the second half.
In a large share of the estates we reviewed, over assigned permission set licenses were the exact line a vendor reconciliation flagged, because assigned count exceeded contracted count. The buyer move is to reconcile assignments to contracted quantities before renewal, and to deprovision the excess. Generous assignment is administratively tidy and commercially expensive.
- Count assignments, not just users. A user can hold several permission set licenses. The bill follows assignments.
- Match assigned to contracted. Pull both numbers and close the gap before the vendor does.
- Deprovision on exit. Removed users often keep entitlements. Strip them in the offboarding flow.
Representative 1000 seat estate. Current USD 2,415k versus optimized USD 1,982k, a USD 433k license reduction. Benchmark scenario, not a quote.
How does renewal cadence compress cost over three years?
It compresses cost when you fix the seat count and the uplift at the one moment the contract lets you change them, which is renewal. Salesforce subscriptions are firm for the term. You cannot drop seats mid term, so the renewal is the single window where the right sized number takes effect for the next three years.
The worked estate below models one representative upper enterprise org. It is a benchmark scenario, not a quote. Every row is seats times list rate times twelve months, and the rows sum to the total.
| Edition | Current seats | List rate | Current annual | Optimized seats | Optimized annual |
|---|---|---|---|---|---|
| Sales Cloud Unlimited | 250 | USD 350 | USD 1,050,000 | 160 | USD 672,000 |
| Sales Cloud Enterprise | 300 | USD 175 | USD 630,000 | 340 | USD 714,000 |
| Service Cloud Enterprise | 250 | USD 165 | USD 495,000 | 210 | USD 415,800 |
| Platform Plus | 200 | USD 100 | USD 240,000 | 150 | USD 180,000 |
| Total | 1,000 | USD 2,415,000 | 860 | USD 1,981,800 |
The optimized estate moves 90 Unlimited users down to Enterprise and removes 140 dormant seats. The result is a USD 433,200 reduction, 17.9 percent against current license spend, before sandbox and storage governance push the total toward the 20 to 30 percent band.
Cadence then multiplies the saving. Doing nothing renews the full count and absorbs the annual uplift, commonly 7 percent. Right sizing and capping the uplift at 4 percent holds the line. The year three gap is where the cadence pays.
Three year spend on the representative estate. Do nothing reaches USD 2,765k by year three; the optimized, capped path holds it to USD 2,144k, a USD 621k year three gap.
Share of provisioned seats removed or right sized across our 2024 to 2025 Salesforce reviews, using login and edition data before any rate negotiation.
Renewal increases absorbed where no uplift cap was written, against the 3 to 4 percent we hold to when the cap and notice window are negotiated up front.
Benchmark ranges: Redress Compliance advisory engagement file, 2024 to 2025.
What is the renewal timeline that protects the saving?
The estates that hit the top of the band started two quarters out, because the only place to bank a right sized count is the renewal, and the notice window closes early. Start late and the count rolls forward at the uplift.
Run the 90 day audit
Pull logins, edition usage, sandboxes, storage, and permission set license assignments. Build the reconciled seat and entitlement count.
Set the position
Decide the retained edition mix, the sandbox and storage controls, and the entitlements to deprovision. Write the target clauses: uplift cap, price hold, notice window.
Negotiate and serve notice
Hold the reconciled count, fix the sandbox fee, and serve any reduction notice inside the window. The order form, not the discount, sets the three year price.
Our Recommendation
Treat Salesforce optimization as an evidence exercise that ends at the renewal table, not a discount conversation. The login data is yours, the right sized count is defensible, and the renewal is the one window where it takes effect.
- Reconcile first, discount last. Strip dormant seats, right size editions, and govern sandboxes and storage before you discuss rate. The reconciled base is what earns the 15 to 30 percent band.
- Win on the order form. Cap the uplift at 3 to 4 percent, fix any percentage based sandbox fee, reconcile permission set licenses, and serve reduction notice inside the window so the count cannot roll forward.
We sit on your side of the table, run the 90 day audit against your org, and negotiate the clauses that set your three year price. We are glad to tie a meaningful part of the fee to delivered value.