Between 30 and 50 percent of Tableau Creator seats went to users who never authored a workbook, and the vCore commit beside them was set without measurement
Both products bill on a unit almost nobody measures. That is not a pricing problem, it is a measurement problem, and it means the savings sit in your own activity reports rather than in the discount ask.
Prepared by Redress Compliance · August 17, 2026 · Salesforce advisory. 20 to 30 Salesforce data platform reviews advised, 2024 to 2025.
Executive summary
Between 30 and 50 percent of Creator seats went to users who never authored a workbook. They belonged on Explorer or Viewer, and the gap between those tiers is large enough that the mix decides the bill.
Estates carried 15 to 35 percent more vCores than their running integrations needed. Because nobody measured live utilization, so the commit was set blind and then renewed against itself.
Usage evidence moves a quote further than a discount ask. Seat activity reports and vCore utilization data are arguments a rep cannot answer, where a request for a better rate is one they answer daily.
Folding the data platform line into the wider Salesforce renewal cost 10 to 20 percent. It lost its own benchmark, and a line nobody benchmarks separately rises unchecked.
Why does the Tableau seat mix decide the bill?
Tableau prices on three seat types and the gap between them is large. Creator authors data sources and workbooks, Explorer edits and explores existing content, and Viewer consumes published dashboards.
Rollouts routinely grant Creator to everyone to avoid access tickets. That convenience is invisible on day one and compounds at every renewal afterward.
| Seat | What it is for | Common allocation | Right sized allocation |
|---|---|---|---|
| Creator | Authors data sources and workbooks | Assigned broadly by default | Builders only |
| Explorer | Edits and explores existing content | Underused | Working analysts, matched to edit activity |
| Viewer | Consumes published dashboards | Underused | The majority of users |
Reclassifying idle Creators to Explorer or Viewer is usually worth more than the headline discount. Between 30 and 50 percent of Creator seats sat with users who never authored a workbook. That is a tier decision made once, at rollout, for a reason that had nothing to do with cost, and then renewed unexamined every year since.
What makes the vCore commit so hard to defend?
MuleSoft bills on vCores, a capacity unit most buyers cannot tie back to a specific running integration. The commit therefore gets set blind and then renewed against its own history.
The detail of that metric is covered at the MuleSoft licensing pillar. What matters alongside Tableau is that both lines share one failure: the billing unit is not something anyone routinely measures.
- Pull live vCore utilization before the renewal, not the provisioned figure, since the gap between them ran 15 to 35 percent.
- Tie each vCore to a named running integration, because a capacity unit with no owner is a capacity unit nobody will defend cutting.
- Cap growth terms on both metrics before signing, since true ups on seats and vCores are where unbudgeted cost appears.
- Keep the data platform line separately benchmarked, even when it sits inside a wider Salesforce agreement.
The Salesforce licence optimisation guide
Seat tier reclassification, the usage evidence that moves a quote, and the growth caps to set before a data platform line folds into the wider renewal.
Get the brief →What 20 to 30 data platform reviews showed
Across roughly 20 to 30 Salesforce data platform reviews Morten Andersen advised on between 2024 and 2025, the overspend sat in the same two places every time, and both were measurement failures rather than pricing failures.
Seat inflation came first. Between 30 and 50 percent of Creator seats went to users who never authored a workbook and belonged on a far cheaper Explorer or Viewer seat. The cause is almost always a rollout decision to grant Creator by default.
vCore blindness came second. Estates carried 15 to 35 percent more vCores than their running integrations needed, because nobody measured live utilization. A commit set against an unmeasured baseline renews against that same baseline indefinitely.
The third pattern is what makes both worse. Where data platform spend folded into the wider Salesforce renewal it lost its own benchmark and rose 10 to 20 percent unchecked, because a line item nobody prices separately is a line item nobody defends.
The common lever is evidence. Seat activity reports and vCore utilization data move a quote further than any discount ask, because they are arguments the rep cannot answer with a concession they were already authorized to give. The wider practice sits at the Salesforce hub.
- Your quote benchmarked against 500,000+ real closed deals, adjusted for size, region, and industry
- Seat tier mix and vCore commit modeled against measured usage rather than provisioned capacity
- True up and growth cap language flagged with the exact quote, the page, and the replacement text
How do you produce the evidence before the quote?
Both metrics have a report behind them, and the work is running it early enough to matter.
Tableau: activity, not assignment
Measure who authored content in the last 90 days rather than who holds a Creator licence. The difference between those two lists is the reclassification opportunity.
MuleSoft: live utilization, not provisioned capacity
Pull actual vCore consumption per running integration. Provisioned capacity tells you what was bought, which is the number you are trying to argue with rather than the one to argue from.
The renewal calendar
Both reviews have to complete before the quote arrives. Evidence produced after a number is on the table is evidence arguing against an anchor that has already set.
The reports to run first
- Tableau content authored per user over 90 days, which separates the builders from the seats assigned to builders.
- MuleSoft vCore consumption per running integration, the only figure that ties the commit to something real.
- A separate benchmark for the data platform line, kept even when it sits inside a wider Salesforce agreement.
What the reviews measured, 2024 to 2025
Across roughly 20 to 30 Salesforce data platform reviews:
Held by users who never authored a workbook and belonged on Explorer or Viewer.
Beyond what running integrations needed, because live utilization was never measured.
Where the data platform line folded into the wider Salesforce renewal it rose 10 to 20 percent unchecked, having lost its own benchmark in the process.
True ups on both metrics are where unbudgeted cost appears, which is why growth terms need capping before signature rather than reviewing after it.
Watch the briefing · 4:44Shrinking a Salesforce EstateWhere the seats actually go, and how a reduction is argued rather than requested.
Your first five moves
- Run the Tableau activity report for the last 90 days and list every Creator seat that authored nothing.
- Reclassify idle Creators to Explorer or Viewer before the renewal quote, since the tier mix is worth more than the discount.
- Pull live vCore utilization per running integration, not the provisioned figure, and tie every vCore to a named owner.
- Keep the data platform line separately benchmarked even inside a wider Salesforce agreement, or it rises unchecked.
- Cap growth terms on seats and vCores at signature. The negotiation practice builds the usage file with you.
Frequently asked questions
How many Tableau Creator seats are misallocated?
Between 30 and 50 percent across the reviews run in 2024 and 2025. Those users never authored a workbook and belonged on a far cheaper Explorer or Viewer seat.
Why does that happen?
Rollouts grant Creator to everyone to avoid access tickets. It is a convenience decision made once, for reasons that had nothing to do with cost, and then renewed unexamined every year after.
What do the three Tableau seats do?
Creator authors data sources and workbooks, Explorer edits and explores existing content, and Viewer consumes published dashboards. The price gap between them is large enough that the mix decides the bill.
Is reclassification worth more than a discount?
Usually. Moving idle Creators down a tier changes the quantity being priced, where a discount only changes the rate applied to a quantity that was wrong to begin with.
What is the vCore problem?
MuleSoft bills on vCores, a capacity unit most buyers cannot tie to a running integration. Estates carried 15 to 35 percent more than they needed, because live utilization was never measured.
How do you measure vCores properly?
Pull actual consumption per running integration rather than the provisioned figure, and give every vCore a named owner. Capacity with no owner is capacity nobody will defend cutting.
What happens inside a wider Salesforce renewal?
The data platform line loses its own benchmark and rises 10 to 20 percent unchecked. A line item nobody prices separately is a line item nobody defends.
When should the usage review happen?
Before the quote arrives. Evidence produced afterward is arguing against an anchor that has already set, which is a materially weaker position than setting the anchor yourself.
Where does unbudgeted cost appear?
In true ups on both metrics. Growth terms on seats and vCores need capping at signature, because both meters move quietly between renewals and neither is reviewed in between.
What is the single strongest lever?
Usage evidence. Seat activity reports and vCore utilization data move a quote further than any discount ask, because they are arguments a rep cannot answer with a concession already authorized.
