How to Build a Data-Driven Account Scoring Model for Outbound Success

‎

About 91% of sales teams we surveyed rely on simple firmographic data in ZoomInfo and spray-and-pray tactics to outbound.

The smartest teams, however, are treating outbound like an engineering problem — building a data-driven account scoring model to prioritize the right leads and create a scalable outbound system.

At Letterdrop, we have just two full-cycle reps. To protect their time while still hitting self-sourced pipeline targets, we’ve developed a system to identify the top 50 accounts every week — automatically.

We use tools like RB2B, Clay, Letterdrop, and Apollo, but the same principles can be applied even if you’re using 6sense or Outreach instead.

‎

Step 1: Pulling in the Right Data

The starting point is collecting accounts of interest. We focus on multiple data sources to cast a wide but intentional net. Here’s what we pull, for example:

• Website Visitors: RB2B identifies about 10–15% of individual visitors who check out key pages like our home page, pricing, or blog. We plan to expand this to Clearbit for account-level insights.
• LinkedIn Engagement: People interacting with our team or posts are flagged by Letterdrop.
• Topic and Influencer Engagement: Prospects engaging with competitors or influencers in our space, also flagged by Letterdrop.
• Cold Lists: Traditional Sales Navigator searches filtered by firmographic data. These accounts have no prior affinity to us, so they’re often harder to convert.

‎

Other signals you might consider include:

• Champions who’ve recently switched jobs (using tools like UserGems or Champify).
• Accounts reviewing competitors on G2.
• Leads from webinar attendees or conference badge scans.

This multi-signal approach gives us a rich pool of accounts to evaluate.

Step 2: Qualifying the Accounts That Matter

This is a lot of accounts. We want to reduce it to down to the accounts that matter.

Tools like RB2B, Letterdrop, and SalesNav have some level of filtering but we want to go a step further to qualify accounts.

We use Clay to find answers to some qualifying questions like:

• Is their CEO active on LinkedIn?
• Are they hiring in sales, marketing, or RevOps?
• Do they already use intent data tools like 6Sense, G2, or Demandbase?
• What’s their sales headcount?

We add points if the account fulfills these criteria and create a weighted score for every account.

Warm leads from Letterdrop and RB2B that are already familiar with us or topics of interest get bonus points because we want to act quickly on them.

Step 3: Enriching Contacts for Outreach

Once we have our top 50, we push them into another Clay table to find the right people at the account. For each account, we:

• Identify decision-makers like Heads of Sales, Marketing, or RevOps.
• Look up email addresses and LinkedIn profiles if we don't have them already from RB2B or Letterdrop.
• Push everything into Apollo, so our BDRs can start outreach across channels.

Step 4: Outbound, but With Context

This isn’t a “spray and pray” operation. We use the supporting context in our outbound messages and hand craft them using templates.

This is a good blend of keeping it human (because no one really wants to receive AI messages) but also scaling.

• Did they recently hire AEs? That’s a signal they might be thinking about spending on efficiency for these new hires.
• Did they engage with a post on LinkedIn? Letterdrop forwards that context so we can connect the dots in our outreach.
• Are they visiting our website or following our company LinkedIn page? They're familiar with our product.

We have a full guide on how to follow up on LinkedIn engagement as a sales trigger, for example.

For warm leads especially, reply rates average 5–10%, far higher than industry standards.

Why This Account Scoring Model Works

The problem with traditional cold outreach is that people are trying to boil the ocean when they reach out to their entire TAM. They miss reaching out to the warm or best fit accounts at the right time. And instead waste resources chasing fake leads.

The system above attempts to fix that by prioritizing people who are already familiar with us, showing signs of being in market, qualifying everyone, and narrowing it down to just the best fit accounts (at least as far as we can tell from publicly available data).

If you’d like to replicate this system, let us know. We’re happy to share our templates or help you customize an account scoring model that fits your team’s needs.