Sales intelligence

How sales intelligence works

Sales intelligence answers three practical questions for a sales team. They are which companies to pursue, who inside them to contact, and what to say and when.

Salesforce defines sales intelligence as the collection of data and insights used to make informed and strategic decisions throughout the sales process 1. IBM describes it as the systematic gathering of information about prospects, customers, competitors and market conditions 2. Both definitions put collection at the center, and the decisions come later.

Salesforce groups the data into five main types 1.

• Firmographics. Industry, company size, locations, revenue and growth stage.
• Contact data. Names, roles, work email, phone and social profiles for the people at an account, the core of B2B contact data.
• Technographics. The software and infrastructure a company runs.
• Sales event triggers. Moments such as expansion, new funding, hiring, layoffs or mergers that may open an opportunity.
• Deal data. Budget, timelines, competitor feedback and history from past and current deals.

IBM adds intent data and competitive intelligence to the list 2. It also lists where the data comes from, including public sources, paid databases, website tracking, social media monitoring and the CRM itself 2.

The data flows into daily work in a few places. It fills fields on CRM records, it builds and filters prospect lists, it feeds lead scoring models and it gives reps the research they need before a call. Most teams blend sources, because no single source covers every account well.

AI has changed how fast this data is gathered. Salesforce's State of Sales research found that 4 in 5 sales teams are experimenting with or have already implemented AI, and it says AI has helped speed up collecting sales intelligence 1. Automated research still needs checks, because a summary of a stale record repeats the error with more confidence.

What makes sales intelligence useful

Having data is easy. Having data a rep can trust on a Monday morning is harder, and four properties decide the difference.

Accuracy comes first. A wrong title or a dead email address costs more than a blank field, because the rep acts on it. Data accuracy should be measured on a sample of records against a known answer, not taken from a vendor's headline figure.

Freshness comes second. People change jobs and companies change size, so records lose value over time. Salesforce notes that event triggers such as funding, hiring and layoffs mark points in time that may create an opportunity 1. An event that is months old usually tells a rep less than one from this week.

Coverage comes third. A source can be accurate for large companies in one country and thin everywhere else. Teams should test coverage on their own target list before buying, and combine sources where one falls short.

Fit with the workflow comes fourth. Data that sits in a separate tool is used less than data that appears on the CRM record. IBM notes that sales intelligence and sales analytics ideally work together, with intelligence supplying the raw data and analytics producing the insight 2.

The case for sales intelligence comes down to time. Salesforce reports that reps spend only 30% of an average week selling, with the rest going to work such as researching prospects 1. Good sales intelligence gives some of that research time back.

Example of sales intelligence in a prospecting run

A RevOps team has a list of 500 company domains from an old CRM export and wants to reach only the companies that run a specific e-commerce platform 3. RichAPI's tech stack recipe runs the cheap fit check first and spends on people only for survivors 3.

Fit filter before contact spend
  clean the domain list         remove tracking junk and duplicates
  web_tech_stack per domain     detected public web technologies
  keep domains that match       the target platform is present
  people_search on survivors    roles that own the platform
  email_finder on the people    misses bill zero
  push to CRM                   with the detection date on the record

The recipe's worked example uses stated counts. Detection on 500 domains at 1 credit each bills 500 credits, including domains where nothing is detected 3. Ninety domains survive, so 82% of the list is removed before any per-contact spend 3. A people search returning 4 contacts per survivor gives 360 results at 0.1 credits each, or 36 credits 3. The email_finder step then costs up to 5 credits per call, and every miss bills zero 3.

The recipe is careful about what the data means. A stack detection shows that a company runs a technology, which makes it a fit signal, and it is not evidence of purchase intent 3. Detection is also limited to technologies visible on public web pages, so an empty result is a limit of the method rather than proof of absence 3.

The team stores the detection date with each account. When the list is reused next quarter, old detections are checked again before anyone writes to the account.

Sales intelligence vs sales analytics

The two are often sold together. IBM draws the line clearly, since intelligence gathers the raw data and analytics turns it into insight about performance 2.