Data prospecting: How to prioritize the right prospects in your CRM

Data prospecting is the practice of using CRM data to determine which potential buyers a sales team should pursue first. Rather than working from a list top-down, data prospecting enables teams to prioritize prospects without resorting to list buying, scraping, or high-volume cold outreach.

Get Started with HubSpot Sales Software for FreeHubSpot’s Smart CRM is one; the scoring model in this guide is built on it. This allows sales teams to easily prioritize the right prospects by applying a simple scoring model in the CRM based on fit, intent, and engagement signals.

In HubSpot’s 2026 State of Sales report, 30% of sales leaders name generating enough qualified leads as their biggest challenge. While a scoring model doesn’t generate more leads, it automatically prioritizes and surfaces them for immediate action.

Table of Contents

What is data prospecting?

Data prospecting uses CRM records to identify likely buyers and rank them by how ready they are to talk. The records are the raw material: Every company property, form submission, page view, tracked email, logged call, and meeting sits on a contact or company record. The CRM’s prospecting model reads those fields to decide which accounts should get a rep’s attention first. That makes the CRM itself the prospect database.

A traditional purchase list provides names and titles at a single point in time. But CRM records accumulate history, so each account can move up or down the priority order as its behavior changes. For example, an account that matched the ideal customer profile in March but hasn’t opened an email since then will rank lower in September than an account that visited the pricing page twice last week.

Why is data prospecting more effective than traditional prospecting?

Cold prospecting starts with outreach and hopes for a signal. Data prospecting starts with the signals already on the record and reserves outreach for accounts that show both fit and interest. With data-backed evidence, sales managers can effectively answer “Why this account today,” and reps can reach out to the best leads right away.

Which signals matter most in data prospecting?

Three kinds of signals matter in data prospecting: Fit, intent, and engagement. A prospect ranks highest when all three signals point the same way.

Fit carries the most weight because it changes the least (an account that’s the wrong size or industry stays wrong no matter how many emails it opens), so fit acts as a gate before the other two are counted.

Engagement is the more reliable of the remaining two because it’s first-party: The CRM recorded the open, reply, and meeting.

Intent from outside sources is the weakest input on its own and works best as a tiebreaker between accounts that already pass on fit.

Fit Signals

Fit signals are the firmographic data and Ideal Customer Profile (ICP) match criteria that show whether an account is a realistic buyer.

  • Company size: Headcount or revenue range, which predicts budget, buying process, and whether the product’s pricing tier makes sense for the account.
  • Industry: The vertical the account operates in, used to match against the industries where the product has already won.
  • Role: The contact’s job title and seniority, which shows whether the person can influence or approve a purchase.
  • ICP match: Whether the account meets the written ideal customer profile as a whole, often stored as a single yes/no or tiered property.
  • Technographic data: The tools the account already runs, which reveal integration fit and whether the product would replace or complement something in place.

Intent Signals

An intent signal indicates that a prospect is researching a relevant problem or solution category, whether or not they have spoken to a rep.

  • Pricing and comparison page visits: Repeat visits to bottom-of-funnel pages, tracked to a known contact or company through the CRM’s website tracking.
  • Topic-specific downloads: A guide or template for the exact problem the product solves, signaling more than a visit to a general blog post.
  • Third-party intent feeds: An outside provider flagging that the prospect is consuming content in the product’s category across the web.

Engagement Signals

Engagement signals record how a prospect interacts with a company’s website, emails, meetings, and sales activity, and they are the most direct evidence that a real conversation is possible.

  • Email opens and replies: Tracked opens show attention; replies show willingness to talk and count for far more.
  • Meetings booked: A scheduled or completed meeting, the strongest single engagement event most models score.
  • Calls logged: Connected calls and their outcomes, recorded on the contact, so the score reflects live conversations.

How to Build a Simple Data Prospecting Score in HubSpot

A data prospecting score in HubSpot is a company score built in the lead scoring tool. Building one takes five steps:

  • Define the ICP and fit criteria.
  • Normalize engagement signals.
  • Add trustworthy intent signals.
  • Weight the score and set thresholds.
  • Turn the score into views, tasks, and alerts.

Company scores are available in Sales Hub Professional and Enterprise. (Contact-level scores need Marketing Hub.) All steps live inside the CRM, so nothing has to be exported, enriched, or purchased before running the model.

Step 1: Define your ICP and fit criteria.

An ideal customer profile (ICP) is a written description of the accounts that close fastest and stay longest. But it comes from closed-won data, not always the biggest names on a wish list.

To define an ICP, sales managers usually pull the last 12 months of won deals and look for commonalities such as industry, headcount band, region, and technology in use. The three or four traits that appear most from the ICP. (Traits that only appear in lost deals become a negative criterion.)

In HubSpot, those traits become property rules in a company score. From Marketing > Lead Scoring, the admin creates a company score, picks fit (or a combination, since engagement arrives in Step 2), and adds one rule per ICP trait. A first pass might look like this:

  • Industry is any of the winning verticals: +15 points.
  • Number of employees is between 50 and 500: +10 points.
  • Country/Region is outside the territories the sales team covers: −20 points.

HubSpot lead scoring builder showing a company score with fit and engagement groups, group limits, and point values

Source

Rules fire independently, so a company that matches industry but not headcount still earns the industry points.

Pro tip: Three to five rules are plenty for a first-fit model. Using more adds unnecessary noise that can make Step 4 harder to weight.

Step 2: Normalize engagement signals.

Normalizing engagement signals means assigning a point value to each engagement activity that reflects the level of interest, then capping the total so volume can’t outrank quality.

Example: 10 email opens from one contact should never outscore a single booked meeting. Without a cap, they will.

Engagement events reach the company record through the CRM’s own tracking. HubSpot’s email tracking software logs opens, clicks, and replies to the contact who received the email. Meetings, calls, and form submissions log the same way. To determine a company score, HubSpot reads the contact activity for a given company, so one score reflects what everyone at the account has done.

Three settings in the score builder do the normalizing:

  • A group limit caps how many points an engagement group can contribute, no matter how many events pile up inside it.
  • A frequency filter lets a rule require an event at least twice, so one stray open earns nothing.
  • Score decay shrinks an event’s points on a schedule of 1, 3, 6, or 12 months, so a meeting from last year stops propping up an account that has gone quiet.

Example of a first engagement group, with points and caps:

  • Email reply from any associated contact: +10 points
  • Meeting booked: +20 points
  • Sales engagement group limit: 30 points, with 50% decay every three months
  • Marketing email open: +2 points
  • Marketing engagement group limit: 10 points

Group limits keep the model honest. An account that replied once and booked a meeting sits at 30. An account that opened 30 newsletters sits at 10 and stays there, no matter what it opens next.

Step 3: Add trustworthy intent signals.

An intent signal is trustworthy when the source is known, and the account could reasonably expect the activity to be seen. In HubSpot, that distinction maps to the data source. First-party intent is activity HubSpot’s own tracking captured in the CRM: A known contact visiting the pricing page or downloading a guide. Third-party intent is data that an outside vendor collected and sold. The first is already on the record and ready to score. The second needs vetting before it earns any points.

First-Party Intent (Captured in HubSpot)

Buyer intent data from the company’s own site is the best place to start because the CRM has already collected it. HubSpot’s website tracking ties page views to a known contact once that person has submitted a HubSpot form or clicked through a tracked marketing email, and only if they accepted the site’s cookies. Without a CRM, a sales team would have to pull web analytics and match visitors to accounts by hand. In HubSpot, page views are tracked on the contact record, and scoring the right pages turns them into an intent signal.

Two rules cover most of it:

  • Pricing page visited at least twice in the last 30 days: +10 points
  • Solution-specific guide downloaded in the last 90 days: +8 points

The two-visit requirement on the pricing page screens out accounts that looked once and moved on. Setting the intent group to decay fully at three months keeps the signal current, since an account that was shopping in June and went silent isn’t shopping in October.

Third-Party Intent (From Outside Vendors)

A third-party intent feed comes from a vendor that tracks which accounts are reading about the product’s category across the web. Before any of that data is used to calculate a score, the vendor needs to say where it comes from and how people can opt out. A feed that clears that bar goes into the same intent group as the first-party rules, at a lower point value and under the same cap.

Step 4: Weight the score and set thresholds.

A rep doesn’t need to see a point total. They need a label on the record that says whether to call this account today. Step 4 is where a raw score turns into that label. Weighting determines how much each signal group contributes to the total; thresholds determine where the labels fall. In HubSpot, both are settings in the same score builder, so the admin sets them once, and every company record picks up the result.

Setting Weights

Without a CRM, weighting means a spreadsheet with a multiplier column and a debate about the multipliers. In HubSpot, the group limits from Steps 1 through 3 are the weights. With an overall score limit of 100, the three groups split like this:

  • Fit group limit: 40 points
  • Engagement group limit: 40 points
  • Intent group limit: 20 points

Fit and engagement carry equal weight because they answer different questions, and an account needs both to be worth a rep’s morning. Intent stays at 20 because it’s the least certain of the three signals, and a cap that size keeps it from outvoting the other two. Negative fit rules, such as the out-of-territory rule from Step 1, pull the total down no matter how engaged the account is, which is the gate doing its job.

Setting Thresholds

Without a CRM, a threshold is a number someone remembers (“call anything over 60”). In HubSpot, thresholds are set in the builder and show up as a label on the record. For a combined score, HubSpot creates three properties for every company: a total score, a fit score, and an engagement score. Fit and engagement scores get High, Medium, and Low labels at ranges the admin sets.

Combined scores get a two-character grade from A1 to C3, where the letter is fit (A is high), and the number is engagement (1 is high). An A1 account fits well and is active. A C1 account is engaged but wrong for the product. An A3 account is a good fit that hasn’t responded yet.

A rep doesn’t need to know that an account scored 67. They need to know it’s an A2, and that A1 and A2 accounts get worked today.

Checking the Score Before Turning It On

Before the score goes live, the admin can test individual records and preview the score distribution across all companies. If 80% of accounts land in A1, the fit rules are too loose. If nothing lands above B2, the engagement points are too low, or the decay is too aggressive. Two or three passes on the point values usually get the shape right.

Step 5: Turn the score into views, tasks, and alerts.

A score only changes behavior once it shows up where reps already work. HubSpot Sales Hub turns prioritized prospects into saved views, tasks, and follow-up workflows, all keyed to the threshold property from Step 4.

Views

This is where reps start the day. A saved view on HubSpot’s companies index page filtered to grade A1 or A2 becomes the daily outreach list. A second view filtered to A3 shows good-fit accounts that haven’t engaged, which is now the list for measured, low-volume outreach. Reps open the view instead of scanning the whole database.

HubSpot companies index page with a saved view filtered to combined score grade A1 and A2

Source

Tasks and Routing

A workflow in HubSpot’s sales automation tools can assign an owner to an unowned company the moment it crosses into A or B, and create a task for that owner with a due date. Routing off the score means the first rep to see a high-scoring account should have it, not whoever happened to be looking.

Alerts

The same workflow can send a notification when a company’s grade improves, so a rep learns that a B2 became an A1 on the day it happened. Alerts should fire on grade changes, not on every point movement, so notifications stay manageable and relevant.

In Sales Hub’s lead management and prospecting workspace, those views, tasks, and alerts sit on a single screen alongside the account’s full history, so a rep can move from the priority list to the record to the next action without switching tabs.

Sales teams that want to turn a scoring model like this one into prioritized views, tasks, and follow-up workflows can get started with HubSpot’s sales software or request a demo.

How to Keep Data Prospecting Clean and Compliant

Keeping data prospecting clean and compliant comes down to five habits, and none of them requires new software.

  • Make the fit properties required. A score can only read what’s filled in. Setting industry, headcount, and region as required fields on company creation means new accounts arrive scorable. A CRM data hygiene pass on existing records retroactively fills the same fields, starting with accounts that already show engagement.
  • Merge duplicates before scoring. Two records for the same company split its engagement between them, and neither crosses a threshold. Deduplication comes before the score is turned on, not after reps notice an account is missing from the A1 view.
  • Let scores expire. Score decay from Step 2 handles engagement. Fit properties need a lighter version: a quarterly check that headcount and industry still match reality, since the company that was 80 people at import may be 300 now.
  • Score preferences, not just activity. A contact who has unsubscribed from marketing email or set a “do not call” preference is telling the sales organization how it wants to be reached. Those properties belong in the model as negative rules or exclusions, and outreach workflows built in Step 5 need to check them before enrolling anyone.
  • Write the rules down. A one-page list of every rule, its point value, and its reason lets a new sales manager understand the model without reverse-engineering it. Edit access to the score stays with one or two admins.

HubSpot’s data enrichment fills in missing company and contact properties from HubSpot’s commercial dataset, third-party providers, and public sources. It’s off by default. HubSpot’s terms place the obligation for notices, opt-out mechanisms, and consent on the account using it, and an individual who opts out of HubSpot’s dataset is flagged on the record.

For a prioritization model, the practical rule is to enrich only the company properties that the fit score needs. Questions about what’s permitted in a given region go to the organization’s legal team rather than a vendor FAQ.

Where AI Helps With Data Prospecting

AI helps with data prospecting in three ways: summarizing an account before a call, recommending the next action, and drafting a first version of outreach for a rep to edit. AI does not replace the scoring model. The model decides who, and AI shortens the work of figuring out what to say.

In HubSpot’s 2026 State of Sales report, 53% of sales leaders say they’ve fully adopted AI in their prospecting workflow. Still, only 9% describe that adoption as optimized, with AI intertwined in workflows that drive meaningful impact. In the same report, 32% say treating AI as a shortcut is the single biggest barrier to getting more value from AI in prospecting.

A scoring model is the opposite of a shortcut. It’s the structure that tells AI which accounts deserve the effort.

Account summaries. Within the sales workspace, HubSpot’s AI generates company research and engagement summaries on the record, so a rep opening an A1 account reads a paragraph instead of a timeline. In the report, 21% of sales leaders list prospecting research as their team’s top AI prospecting tactic.

Signals and next steps. HubSpot’s prospecting agent, part of Agent Hub, monitors buying signals on target accounts and tells reps when to reach out and why. Pointed at the A2 and A3 views from Step 5, it watches the accounts that fit but haven’t moved, which is where a timing signal matters most.

First-pass messaging with human review. The prospecting agent drafts outreach in each rep’s voice, grounded in the account’s signals, and every draft can be reviewed and edited before it sends. The report backs the review step: reps are still expected to make decisions without AI assistance on account prioritization, personalization, and prospect qualification. A fully autonomous mode exists, and the right time to use it is after a rep has read enough drafts to trust the output.

Predictive lead scoring, where AI builds the score from past conversions rather than an admin assigning points by hand, is the next step up from the manual model in this guide. In HubSpot, it’s available for contact scores on Marketing Hub Enterprise. Teams on Sales Hub benefit most from revisiting the manual weights quarterly. For a wider survey of AI sales prospecting approaches and sales prospecting tools, HubSpot’s guides cover the category beyond the CRM.

Frequently Asked Questions About Data Prospecting

Putting the Prospecting Score to Work

Data prospecting comes down to three signals and one decision rule. Fit gates, engagement leads, intent breaks ties, and an account earns a rep’s attention only when the score says so. Built as a company score in HubSpot, the model reads properties and activity the CRM already holds. Nothing has to be bought, scraped, or enriched in bulk.

The five steps in this guide produce a working version in an afternoon:

  • An ICP written from closed-won data
  • Engagement points capped and decaying
  • First-party intent scored on the pages that matter
  • Weights and thresholds that turn a number into a grade
  • Views and workflows that put the grade in front of reps

Quarterly reviews, required fields, and respected preferences are what keep the A1 view trustworthy six months in.

HubSpot Sales Hub is where the model becomes a workflow: Prioritized views on the companies index, tasks routed to the right owner, and alerts when an account’s grade changes. Sales teams ready to put a scoring model to work can get a demo of Sales Hub and build the first version on their own data.

Leave a Reply

Your email address will not be published. Required fields are marked *