Data Cloud is Salesforce’s fastest-growing product — and its most opaque pricing model. Unlike per-user licences, Data Cloud bills by consumption: credits that tick down with every ingestion, query, and segment refresh. This guide decodes the mechanics so your procurement team can forecast, negotiate, and control costs before they spiral.
Salesforce Data Cloud — rebranded as Data 360 in late 2025 — operates on a fundamentally different commercial model from every other Salesforce product. There are no per-user seats. There is no fixed monthly fee that scales linearly with headcount. Instead, Data Cloud charges based on consumption: the volume of data you ingest, the frequency of profile unification, the number of segments you publish, and the compute intensity of your queries. Every action draws down from a finite pool of credits that your organisation purchased upfront.
This model creates three problems that traditional per-seat procurement teams are not equipped to handle. First, costs are unpredictable — a single misconfigured segmentation job can burn through 30% of an annual credit allocation in weeks. Second, the pricing mechanics are opaque — credit multipliers vary by action type, by batch versus streaming mode, and by sandbox versus production environment. Third, unused credits do not roll over at contract expiry, creating a use-it-or-lose-it dynamic that incentivises overconsumption rather than efficiency.
This guide breaks down every component of Data Cloud’s pricing structure, maps the credit consumption rates for each major feature, and provides the commercial strategies that Redress Compliance’s advisory team uses to help enterprise clients forecast, negotiate, and control Data Cloud costs.
Data Cloud’s total cost comprises three distinct layers: consumption credits, data storage, and premium add-ons. Understanding how each layer is priced and metered is the foundation of any credible cost forecast.
Credits are the core pricing mechanism. Every action performed in Data Cloud — ingesting records, running identity resolution, refreshing segments, executing queries, publishing calculated insights — consumes credits from a prepaid pool. Salesforce sells credits in bundles of 100,000, with a list price of $500 per 100,000 credits. Credits are fungible: a single credit can be spent on any supported action, and all credits now draw from a unified pool (following the September 2025 simplification that consolidated the previously separate Data Services and Segmentation & Activation credit categories into one).
Credits are consumed each time a process runs, not just during initial configuration. An identity resolution ruleset that processes 5 million profiles consumes credits every time it refreshes — whether that refresh was triggered manually, by schedule, or by an upstream data change. A calculated insight that runs nightly across 15 million rows consumes credits every single night. This recurring consumption is the primary reason organisations underestimate Data Cloud costs: they model the initial setup but fail to account for ongoing operational credit burn.
Unused credits expire at the end of the contract term. There is no automatic rollover. If your organisation purchases 5 million credits annually and consumes only 3 million, those 2 million surplus credits are lost. Conversely, if you exceed your allocation, overage charges apply — typically at a higher per-credit rate unless you have negotiated overage pricing in advance. This makes accurate consumption forecasting a commercial imperative, not a technical nice-to-have.
Storage is charged separately on a flat rate per terabyte per month. Data Cloud storage is independent of the credit system — you pay for it regardless of whether you actively process the stored data. Storage accumulates as you ingest records from external sources, create unified profiles, and generate derived objects (calculated insights, segments, data graphs). Service operations in particular generate high storage volumes: case records, chat transcripts, email threads, and event streams from web and mobile interactions.
The critical detail for procurement is that storage is billed on actual stored volume, not on the number of records. A single customer profile that includes rich behavioural data (purchase history, web interactions, email engagement, mobile app events) can consume significantly more storage than a simple CRM contact record. Organisations migrating from traditional CRM-only deployments consistently underestimate their Data Cloud storage requirements by 40–60%.
Data Cloud offers optional premium features that carry their own fixed-cost pricing, separate from both credits and storage.
| Add-On | List Price | Use Case |
|---|---|---|
| Ad Audiences | ~$2,400/audience/year | Activate segments to Google Ads, Meta, Amazon Ads |
| Real-Time Profiles | ~$750/10,000 profiles | Maintain profiles with sub-second real-time updates |
| Data Cloud One | $60,000/instance | Connect additional Data Cloud instances |
| Data Spaces | Contract-dependent | Partition data across business units |
| Platform Encryption | Requires Salesforce Shield | Customer-managed encryption keys for Data Cloud |
The Ad Audiences add-on is the most common surprise cost for marketing teams deploying Data Cloud. Segmentation and activation to internal channels (email via Marketing Cloud, Sales Cloud actions) consume only credits. But activating the same segment to an external advertising platform requires the Ad Audiences add-on — a separate, per-audience annual fee that sits outside the credit model entirely.
The relationship between a Data Cloud action and its credit cost is governed by multipliers — published rate cards that define how many credits are consumed per million rows processed for each action type. These multipliers are the single most important variable in any Data Cloud cost forecast, and they vary dramatically by feature.
| Action Type | Credit Multiplier (per 1M rows) | Cost Impact |
|---|---|---|
| Batch Data Ingestion (external sources) | ~2,000 | Low–moderate |
| Batch Data Ingestion (Salesforce connectors) | 0 (free since Aug 2025) | None |
| Streaming Data Ingestion | ~5,000–10,000 | High |
| Identity Resolution (profile unification) | ~50,000–100,000 | Very high |
| Calculated Insights | ~1,000 | Moderate |
| Segmentation (batch) | ~500–1,000 | Low–moderate |
| Activation (to internal channels) | ~500–1,000 | Low–moderate |
| Queries / Data Explorer | ~100–500 | Low (but accumulates) |
Identity resolution is the single most expensive feature in Data Cloud. When the platform matches, links, and merges records from multiple data sources into unified customer profiles, it performs compute-intensive operations across every row in every participating data stream. A profile unification run across 5 million records can consume upwards of 250,000–500,000 credits in a single execution. If that ruleset refreshes daily, annual credit consumption from identity resolution alone can reach 90–180 million credits — far exceeding what most organisations initially budget for.
The implication is clear: organisations must make deliberate architectural decisions about which data sources participate in identity resolution, how frequently rulesets refresh, and whether incremental (changed-records-only) processing can replace full refreshes. These are not technical decisions alone — they are directly tied to credit consumption and therefore to commercial spend.
“We have seen organisations burn through 30% of their annual credit allocation in six weeks from a single misconfigured identity resolution ruleset. The credit model punishes trial-and-error — you must design your data architecture before you start processing.”
In September 2025, Salesforce implemented the most significant set of Data Cloud pricing changes since the product’s launch. These changes simplified the model in several important ways, but also introduced new dynamics that procurement teams need to understand.
Structured data from four Salesforce connectors is now ingested at zero credit cost. The four connectors are: Salesforce CRM (Sales and Service Cloud), Marketing Cloud Engagement, Marketing Cloud Personalization, and Commerce Cloud. Previously, ingesting Salesforce data consumed credits like any other source. This change removes a significant upfront cost barrier and means organisations can now ground Agentforce AI agents in their structured Salesforce data without consuming credits.
What it does not cover: Ingestion from non-Salesforce external sources (databases, data warehouses, third-party APIs, Snowflake, BigQuery) still consumes credits at the standard batch or streaming multiplier rates.
Four separate credit categories consolidated into one fungible credit. Previously, Data Cloud maintained separate credit pools for Data Services, Segmentation & Activation, sandbox, and production. Organisations had to forecast consumption across each pool independently — a process that frequently resulted in surplus credits in one pool and shortfall in another. The unified credit model eliminates this waste, allowing a single credit to be used for any action across any environment.
Sandbox discount: Sandbox environments now consume credits at a 20% discount compared to production, encouraging testing and development without excessive cost penalty.
Granular usage-tagging insights at the feature level. The Digital Wallet — Salesforce’s built-in consumption monitoring tool — now shows exactly which features, processes, and data streams are driving credit consumption. Organisations can build custom reports, set consumption threshold alerts, and track spending trends in near real-time. This is a significant improvement over the previous model, where customers had to request consumption reports from their Account Executive.
The separate Segmentation & Activation add-on licence has been retired. Effective September 2025, all Data Cloud licences include access to segmentation and activation features without requiring an additional licence purchase. Segmentation and activation actions now simply consume credits from the unified pool, rather than requiring a separate entitlement.
Data Cloud is available through several SKUs (stock-keeping units), each targeting a different entry point. The naming has evolved rapidly — the product was known as Customer Data Platform, then Data Cloud, and is now being marketed as Data 360 as of late 2025.
| SKU | List Price | Included Credits | Target Audience |
|---|---|---|---|
| Data 360 Provisioning (Everywhere) | $0 | Limited (varies by base product) | Existing Salesforce customers needing Data Cloud functionality for other products |
| Data 360 Starter | $60,000/year | Initial credit allocation + storage | Organisations with dedicated Data Cloud use cases |
| Included in Agentforce 1 Editions | Part of $550/user/month | 2.5M Data Services Credits/org/year | Enterprises deploying Agentforce at scale |
| Additional Credit Packs | $500/100K credits | As purchased | Top-up when allocation is exceeded |
The Data 360 Provisioning SKU ($0) is automatically enabled for Salesforce customers whose other products depend on Data Cloud functionality (for example, Marketing Cloud Growth Edition, Agentforce add-ons). It provides limited credits and storage — enough for basic operations but not for enterprise-scale data unification or segmentation. Organisations that outgrow this allocation must purchase the Data 360 Starter SKU or additional credit packs.
The Data 360 Starter at $60,000 per year is the standard paid entry point. It includes a base allocation of credits and storage, with the specific quantities varying by negotiation. Procurement teams should treat the Starter SKU as a floor, not a ceiling: most enterprise deployments require additional credit packs within the first 6–12 months as usage scales beyond initial projections.
Accurate forecasting requires modelling every process that will run in Data Cloud, its frequency, the volume of records it will process, and the applicable multiplier. The following worked examples illustrate how credit consumption scales across different deployment profiles.
Batch ingestion (external): 2M records × multiplier 2,000 per 1M = 4,000 credits per refresh. Weekly = 208,000 credits/year.
Identity resolution: 2M profiles × multiplier ~50,000 per 1M = 100,000 credits per run. Weekly = 5,200,000 credits/year.
Segmentation: 2M profiles × 5 segments × multiplier ~500 per 1M = 5,000 credits per refresh cycle. Weekly = 260,000 credits/year.
Total estimated annual consumption: ~5.7 million credits. At $500 per 100K, that is approximately $28,500 per year in credit costs alone, plus the $60,000 Starter SKU and storage fees. Identity resolution accounts for 91% of the total credit spend.
Streaming ingestion: 500K events/day × multiplier ~10,000 per 1M = 5,000 credits/day = 1,825,000 credits/year.
Identity resolution (daily): 10M profiles × multiplier ~75,000 per 1M = 750,000 credits per run × 365 = 273,750,000 credits/year.
Calculated insights (daily): 10M rows × multiplier 1,000 per 1M = 10,000 credits/day = 3,650,000 credits/year.
Total estimated annual consumption: ~279 million credits. At list price, that exceeds $1.39 million per year in credit costs — before storage, add-ons, or the base SKU. The daily identity resolution schedule is the overwhelming cost driver. Shifting to weekly resolution would reduce that single line item by 85%, saving over $1.18 million annually.
These examples demonstrate why Data Cloud procurement cannot be delegated to technical teams alone. The choice between daily and weekly identity resolution, between streaming and batch ingestion, between full and incremental refreshes — these are million-dollar architectural decisions that must be made jointly by IT, marketing, and procurement. If you are entering a Data Cloud negotiation, these scenarios should form the basis of your commercial model.
Data Cloud’s consumption model creates negotiation dynamics that are fundamentally different from seat-based Salesforce products. The following strategies reflect approaches that Redress Compliance deploys across enterprise Data Cloud engagements.
The standard Data Cloud contract charges overages at list price ($500 per 100K credits) unless you have negotiated a discounted overage rate. Given the inherent unpredictability of consumption-based models, overage clauses are not hypothetical — they are near-certain. Negotiate a tiered overage rate (e.g., 15–25% below list) and ensure the contract specifies automatic notification when consumption reaches 75% and 90% of the allocated pool. Without these protections, organisations routinely discover overages only when the invoice arrives.
First-year Data Cloud deployments almost universally underestimate credit consumption. Configuration, testing, identity resolution tuning, and iterative segment development all consume credits — including in sandbox environments (at the 20% discounted rate). Procurement teams should model their expected consumption, then add a 25–40% buffer to the initial credit purchase. The cost of purchasing surplus credits upfront (at a negotiated volume discount) is consistently lower than paying overage rates mid-term.
Salesforce’s Flex Agreement framework allows organisations to convert unused human (per-seat) licence entitlements into Data Cloud credits, and vice versa. This is particularly valuable for enterprises that are simultaneously deploying Agentforce (which includes Data Cloud credits) and standalone Data Cloud. If your organisation holds both seat-based and consumption-based Salesforce contracts, negotiate the right to reallocate value between them at renewal. This prevents the common scenario where credits expire unused in one contract while another contract faces overages.
Credit multipliers are published on Salesforce’s rate card pages, but they are not guaranteed to remain static. Salesforce has modified multipliers in previous rate card updates, and a change in the identity resolution multiplier — even a modest increase — can have an outsized impact on annual spend. Negotiate explicit rate-card commitments in the contract: the multipliers applicable at contract signing should be locked for the full term. If Salesforce adjusts multipliers mid-term, your organisation should not bear the increased cost against previously purchased credits.
Salesforce account executives frequently attempt to bundle Data Cloud into the broader CRM renewal (Sales Cloud, Service Cloud, Experience Cloud) as a single proposal. This bundling obscures the individual cost components and makes it difficult to negotiate Data Cloud terms independently. Insist on separate line items and, where possible, separate contract schedules for Data Cloud. This preserves your ability to renegotiate or right-size Data Cloud independently of the core CRM at the next contract renewal.
The most effective way to control Data Cloud costs is to optimise credit consumption at the architectural level. The following practices consistently deliver the largest savings across enterprise deployments.
Identity resolution is the single largest credit consumer. Reduce its cost by: ingesting only active profiles (exclude records inactive for more than 12–18 months), applying data filters at the data lake object level to limit records entering the unification pipeline, validating match rules against a sample dataset before running against the full population, and using incremental processing (changed records only) rather than full refreshes. Organisations that implement all four practices typically reduce identity resolution credit consumption by 50–70%.
Streaming ingestion consumes 3–5x more credits than batch ingestion for the same data volume. Before enabling streaming for any data source, challenge the business requirement: does this use case genuinely require sub-minute data freshness, or would hourly or daily batch updates deliver equivalent business outcomes? In most enterprise deployments, fewer than 20% of data streams actually require real-time ingestion. The remaining 80% can run in batch mode at a fraction of the credit cost.
Segments can be configured to refresh every 1, 4, 12, or 24 hours. Marketing teams default to the most frequent schedule available, but the business impact of hourly versus daily segmentation is often negligible for batch campaign use cases. Moving a segment from hourly to daily refresh reduces its credit consumption by up to 96%. Review every active segment’s refresh schedule quarterly and deactivate segments that are no longer associated with live campaigns — dormant segments still consume credits on their configured schedule.
Batch data transforms within Data Cloud consume credits. If your organisation has existing ETL infrastructure (MuleSoft, Informatica, dbt, custom pipelines), perform data cleansing, deduplication, and transformation before ingesting into Data Cloud. This reduces both the volume of records ingested and the compute cost of in-platform transformations. The savings are particularly significant for large-scale data migrations, where cleaning data externally before a one-time bulk load can save hundreds of thousands of credits.
Data Cloud is not a standalone product in Salesforce’s strategic vision — it is the data foundation for Agentforce. Every Agentforce agent that retrieves customer context, grounds responses in historical data, or personalises interactions draws from Data Cloud. This dependency means that organisations deploying Agentforce must factor Data Cloud credit consumption into their AI cost model.
The Agentforce 1 edition ($550/user/month) includes 2.5 million Data Services Credits per org per year. For small-to-mid-scale AI deployments, this allocation may be sufficient. For enterprises running agents across sales, service, and marketing at scale, 2.5 million credits will be consumed within months. Additional credit packs then become a recurring operational expense that compounds the already-premium Agentforce per-user fee.
The strategic advice for most enterprises is to model Data Cloud credit consumption for Agentforce use cases before committing to Agentforce licences. The base Agentforce licence cost is visible; the Data Cloud credit tail is where budgets overrun. Use a total cost of ownership analysis that includes both per-user and consumption-based costs to build a credible business case.
Data Cloud’s consumption model demands a governance discipline that most organisations do not have for seat-based products. Unlike per-user licences where the cost is fixed and predictable, Data Cloud costs can spike without warning if a new data stream is added, a refresh schedule is changed, or a query runs against an unexpectedly large dataset.
The Digital Wallet is the primary monitoring tool. It provides near real-time visibility into credit consumption by feature, data stream, and process type. Organisations should configure consumption threshold alerts at 50%, 75%, and 90% of the annual allocation. Additionally, Salesforce exposes Tenant Billing Events via Platform Events, enabling teams to build custom monitoring dashboards, automated alerts, and even automated responses (such as pausing non-critical processes when consumption exceeds a defined threshold).
Governance responsibility should be assigned to a specific role — typically the Salesforce Centre of Excellence lead or IT Asset Management team. This role should review weekly consumption reports, approve any new data streams or process schedules that will increase credit consumption, and conduct quarterly reconciliation between actual consumption and the contracted credit allocation. Without this discipline, credit overruns become inevitable. Integrating Data Cloud monitoring into your broader Salesforce licence optimisation practice is essential.
Redress Compliance provides independent Salesforce advisory — no vendor partnerships, no referral fees. We help enterprises forecast Data Cloud credit consumption, negotiate consumption-based contracts, and implement governance frameworks that prevent cost overruns.