Who to hire for Datorama on Salesforce

Of all the products Salesforce acquired for marketing, Datorama is the one where the hiring question is hardest to answer from a CV. It arrived as a standalone analytics platform, took on a new name, and now sits next to a second, newer Salesforce analytics product with a similar name. Add the fact that marketing analytics projects live or die on whether people trust one number, and you can see why logo slides are weak evidence here. This page explains how to work out which product you have, what skills the work needs, how to run a hiring session that reveals competence, and which red flags belong to this product. It names no winner. Project pricing is on Datorama implementation cost.

#Which product are you hiring for?

The rebrand record states the history: "Datorama, part of the suite of tools that formed CRM Analytics, became Marketing Cloud Intelligence and its reporting tool, Datorama Reports for Marketing Cloud, became Intelligence Reports for Engagement." That is the first product: Marketing Cloud Intelligence, the platform you may still hear called Datorama.

Salesforce's own analytics page then draws a distinction that matters for every brief: "Marketing Intelligence is Salesforce's new marketing analytics product built natively on the Salesforce Platform." And it describes the older platform this way: "Marketing Cloud Intelligence (formerly Datorama) is a separate, enterprise-scale platform for more complex, multi-source analytics needs."

So there are two products. Marketing Cloud Intelligence is the enterprise-scale platform for complex, multi-source analytics. Marketing Intelligence is the newer product built on the core platform. Their skills overlap but are not identical. A partner who builds on the Salesforce Platform is not automatically fluent in Marketing Cloud Intelligence's own data model, and the reverse is true as well.

Before you brief anyone, write down the product named on your contract, the number of data sources feeding it, and the report your executives actually read. If you cannot name all three, your first spend should be a fortnight of internal discovery, not an agency.

#The skills the work really needs

Marketing analytics platforms reward people who think like data engineers and communicate like analysts. For Marketing Cloud Intelligence the list is:

• Data ingestion. Setting up and troubleshooting data streams from advertising platforms, web analytics, email, CRM and files, including API-based and scheduled loads.
• Mapping and harmonisation. Matching fields from many sources to a shared model, so that spend, clicks and revenue from different platforms line up. This is where most arguments about numbers start.
• The data model. Dimensions and measurements, and the discipline to stop teams adding a new measure for every request.
• Calculated measures. Derived metrics such as cost per acquisition or return on spend, which must be defined once and used everywhere.
• Dashboards and reports. Designing views people read, with filters and access that match the organisation.
• Governance. Naming conventions, ownership, change control and a record of what each source means.

Salesforce's Engagement credentials are well documented on Trailhead, but while preparing this page we could not confirm a current credential specific to Marketing Cloud Intelligence. Treat any badge on a CV as a conversation starter and test the person instead.

#The working-session test

Because credentials are a weak guide here, use a working session. It is the single most useful piece of hiring practice for this product.

Book two hours with the candidate consultant, the marketer who trusts a particular report, and an analyst who doubts it. Give the consultant read access to the workspace, and ask them to trace one headline number back to its source: from a dashboard widget, to a calculated measure, to a mapped field, to a data stream, to the original platform.

Watch for four things. Do they find the chain quickly? Do they ask which platform is the source of truth for spend? Do they spot a mapping that would double count? Do they explain it in words the marketer understands?

A candidate who passes this has done the work. A candidate who talks about "a single source of truth" without opening anything has not. It costs a morning and spares you from a bad month.

#Why marketing numbers disagree

Most analytics disputes are not bugs, they are definitions. A partner who knows this will spend their first week on definitions, so it helps to know the usual suspects before you hire.

Time zones come first. An advertising platform reports in the account's time zone, your web analytics in another, and your CRM in a third. A day's spend can land on different days in different tools, and weekly totals drift.

Currency is next. Spend in one currency and revenue in another need a stated conversion rate and a stated date. If the workspace converts at load time in one stream and at report time in another, totals will differ in ways nobody can explain.

Attribution windows follow. One platform claims a conversion within a long window after a click, another within a short one, and the CRM records only the final sale. None of them is wrong. They answer different questions, and a dashboard that puts them side by side without labelling the windows invites argument.

Then there is lag. Some platforms revise numbers for days after the fact. A report run on Monday for last Friday will differ from the same report run a week later. Ask the partner how they treat provisional data.

Last is duplication. When two streams carry the same spend, for example a direct connection and an agency export, totals double. This is a mapping problem, and it is easy to introduce in good faith.

A consultant who can talk through these without notes has done the work. One who answers with a platform feature name has not.

#A first engagement worth buying

Start with a trust audit of one report, over a few weeks. The scope is deliberately narrow: the report the executive team uses, and the numbers on it. The partner documents each measure's definition, its sources, its mapping, and any known gap. They then reconcile the figures with the source platforms for a sample period, and list the differences with causes.

The outputs are a measure dictionary, a reconciliation, a list of fixes ranked by the size of the discrepancy, and a recommendation on whether to repair, restructure or retire parts of the workspace. The business owner of the number signs it off.

Only after that should you scope additions, such as new sources or a new dashboard. If you are considering the newer Marketing Intelligence product, the audit gives you the facts for the comparison, and should I migrate from Datorama sets out the decision.

#Red flags specific to this product

Redrawing before reconciling. A partner who offers to rebuild your dashboards in another tool before checking whether the numbers are right will reproduce your errors faster.

One product assumed. If a proposal never says whether it is for Marketing Cloud Intelligence or the newer Marketing Intelligence, the author has not read your contract.

A "single view" promise. Single views are a design outcome, not a feature. Ask which sources, what keys and what refresh schedule.

No measure dictionary. If the plan does not include written definitions, every future argument starts from zero.

Core-platform thinking. Proposals that talk about objects, flows and permission sets are describing a different system. Marketing Cloud Intelligence has its own structure.

Silence on cost of data. Volume, refresh frequency and the number of sources shape licensing and performance, so a partner should raise them early. Pricing detail sits on the Datorama licence page.

#Ten questions for the first call

1. Which Intelligence product do you think we have, and how did you check?
2. Walk me through tracing one dashboard number back to its source.
3. How do you handle two platforms reporting different spend for the same campaign?
4. How do you decide which mapping becomes a shared dimension?
5. Show me how you document a calculated measure so an analyst can challenge it.
6. What is your policy on adding a new measure at a marketer's request?
7. How do you manage access so agencies see only their own data?
8. Which data streams fail most often, and how do you monitor them?
9. How would you compare our workspace with the newer Marketing Intelligence product, fairly?
10. What would you retire from our workspace?

A good answer to number three is practical and specific, covering time zones, currency, attribution windows and platform definitions of a click. A poor answer is a slogan.

#Checking references and evidence

Ask a reference which numbers changed after the project and who was surprised. A well-run analytics project does surface uncomfortable differences, and a reference who says everything stayed the same is either lucky or incurious.

Ask for a redacted measure dictionary or mapping document. Analytics partners should have these and should be proud of them.

Ask whether the previous client's analysts still use the workspace a year later, and whether anyone has added measures without breaking old ones. Adoption and stewardship are the real tests.

Check the individual, not the firm. Marketing analytics projects depend heavily on one or two people. Ask to speak with the named lead and see their previous workspaces described in detail.

#Commercial models and the statement of work

The audit suits a fixed fee. New data sources suit a unit-based price per source, with a cap and a definition of done that includes reconciliation. Ongoing dashboard work suits a retainer with a monthly allowance and a backlog you control.

The statement of work should name the product, the workspace and the data sources in scope. It should list deliverables as artefacts: a measure dictionary, a mapping document, a reconciliation, a handover recording. It should state who owns each metric definition and who approves a change. Include acceptance tests that compare figures with source platforms for an agreed period. State plainly what is out of scope, including any rebuild in another tool, and require your sign-off before any data stream is deleted or remapped.

The commercial history is on what happened to Datorama, and the connectors are in integrating Datorama with Salesforce. For the newer product, see Salesforce's page on Marketing Intelligence.

#When you can skip a partner

If a competent analyst already owns the workspace and the request is a new filter, a refreshed dashboard or a source that has an existing connector, do it internally. Skip an agency if the real issue is that two teams disagree about a definition, because no partner can settle that for you. Hire help when sources are numerous, numbers do not reconcile, or you are deciding between the older and newer products and need a fair comparison.

If you do compare the two products, insist on a like-for-like test. Choose three questions your marketers genuinely ask, such as which channel produced the cheapest qualified lead last quarter, and answer each in both products using your own data for the same period. Record how long each took, how many manual mappings were needed, and whether the answers agreed. A comparison built on a vendor demonstration will always favour the demonstration. One built on your data tells you what you will actually live with, and it gives the eventual decision a paper trail that finance and marketing can both accept.