Account Intent Data Quality & Scoring Guide

Introduction

ABM success depends on intent data quality. Bad intent data wastes time chasing uninterested accounts. Good intent data surfaces high-probability deals. This guide breaks down how to evaluate intent data sources, understand scoring accuracy, and choose which signals matter for your business.

Learn more about account-based marketing strategies.

Types of Intent Data

First-Party (Your Own Data)

What it is: Signals from accounts interacting with your properties.

Examples: - Website visits and content consumption - Email opens and clicks - Demo requests or form submissions - Product trial signups - Ad clicks and landing page time

Accuracy: 95%+ (your own data is the most reliable) Freshness: Real-time Cost: Included in marketing platform (HubSpot, Marketo)

Best for: Measuring engagement with accounts you already know about

Limitation: Doesn't help identify unknown accounts in-market

Second-Party (Partner Data)

What it is: Intent signals from your vendors' platforms.

Examples: - LinkedIn engagement signals (LinkedIn seeing who engages with your company) - Slack mentions of your product - GitHub stars or commits to your libraries - Community forum discussions mentioning you

Accuracy: 80-90% (mediated through partners) Freshness: Weekly to daily Cost: Usually bundled with platform

Best for: Discovering accounts actively engaging with your brand

Limitation: Limited to accounts already aware of your brand

Third-Party (Bombora, G2, etc.)

What it is: Signals aggregated from websites you don't own.

Examples: - Bombora: 30k+ B2B website consortium tracking purchase signals - G2: Research activity on G2 listings - Madison Logic: IP-based website visit tracking - Demandbase: Proprietary intent network

Accuracy: 60-85% (probabilistic, not deterministic) Freshness: Daily to weekly Cost: $1k-5k per month

Best for: Discovering accounts in-market who don't know about you yet

Limitation: Probabilistic signals, not deterministic. High signal = account was researching category, not necessarily buying from you.

---

Bombora: The Intent Data Standard

How Bombora Works

Bombora aggregates purchase signals from 30k+ B2B websites (company sites, Gartner, Capterra, etc.).

Signal types: - Buying signals: Page views, downloads, demos on B2B vendor sites - Company research: Earnings calls, regulatory filings, news mentions - Role-based signals: Specific titles viewing product comparisons - Technographic signals: News about tool adoption, company tech stack

Scoring: 0-100 (higher = more active buying signals in last 30 days)

Example: - 20-30: Company visited a competing product demo page - 50-60: Company visited 3+ ABM product comparison pages - 70-80: Company downloaded 5+ ABM resources, viewed pricing - 90-100: Company visited ABM vendor sites 10+ times in past 2 weeks

Bombora Accuracy

• Precision (if we say account is high-intent, is it?): 60-70%
• Recall (do we catch all high-intent accounts?): 70-80%

In English: If Bombora flags 100 accounts as high-intent, 60-70 are actually evaluating. We probably miss 20-30% of accounts truly evaluating.

Best for: Identifying accounts actively researching your category

Not for: Identifying accounts passively aware but not yet in-market

Cost: $1.5k-3k/month for 50-100 accounts

G2: Research Activity Intent

How G2 Works

G2 tracks who's researching products on their platform.

Signal types: - Product page views: Who viewed ABM products on G2 - Comparison views: Who compared ABM tools (Demandbase vs. Abmatic AI, etc.) - Review reads: Who read customer reviews - Category research: Who browsed the ABM category

Scoring: 0-100 (research activity intensity)

Example: - 20-30: Company viewed 1-2 ABM products - 50-60: Company viewed 5+ ABM products, read reviews - 70-80: Company viewed products, read reviews, downloaded comparison report - 90-100: Company viewing ABM products multiple times per week

G2 Accuracy

• Precision: 50-65% (many research browsers don't buy)
• Recall: 85-95% (G2 catches most accounts researching)

In English: G2 is good at finding accounts in research phase, but many are just browsing, not actively buying. But if an account's not on G2, they're probably not researching (high recall).

Best for: Identifying accounts researching your product category

Not for: Prioritization (too many false positives)

Cost: $2k-5k/month depending on volume

First-Party Intent: Website & Email

How First-Party Intent Works

You track: - Who visits your website (and which pages) - Who opens your emails - Who clicks your ads - Who attends your webinars

Scoring: 0-100 based on: - High-value page views (pricing, comparison, demo) = 80-100 - Mid-value page views (blog, case study) = 40-60 - Low-value page views (homepage) = 10-20 - Email engagement = 5-10 per email open

Example: - Company visits pricing page + demo page = 90 intent score - Company downloads case study + views comparison = 70 intent score - Company views blog post once = 20 intent score

First-Party Accuracy

• Precision: 85-95% (accounts visiting pricing are seriously considering)
• Recall: 10-20% (only catch accounts aware of you)

In English: First-party intent is highly accurate, but only for accounts who already know about you. You miss 80-90% of market.

Best for: Prioritizing known accounts that are actively evaluating

Not for: Discovery (only works for known accounts)

Cost: Included in HubSpot, Marketo, Clearbit

Combining Intent Sources: The Hybrid Model

Most effective ABM uses multiple intent sources:

1. Discovery phase: Use Bombora or G2 to identify accounts in-market
2. Engagement phase: Send them content
3. Intent confirmation: Measure first-party intent (website visits, email engagement)
4. Qualification: Sales team evaluates fit and buying committee

Example workflow:

• Week 1: Bombora flags Company X as high-intent (score 75+)
• Week 2-3: Marketing sends email + ads to Company X
• Week 2-4: Company X visits your website, pricing page, demo request form (first-party intent now 85+)
• Week 4: Sales team calls; confirms buying committee is evaluating
• Week 5+: Dedicated sales engagement

Cost of hybrid model (50 accounts): - Bombora: $2k/month - First-party (included in HubSpot): $300/month - G2 (optional): $2.5k/month - Total: $2.3k-4.8k/month

Intent Data Quality Red Flags

Watch out for:

1. Claiming 100% accuracy - Intent data is probabilistic. Any vendor claiming perfect accuracy is lying.

2. Not refreshing signals frequently - Intent signals decay. If Bombora data isn't updated daily, it's stale.

3. Mixing up correlation with causation - "95% of customers showing these signals converted" doesn't mean 95% of accounts showing these signals will convert. (Many accounts show signals but don't buy.)

4. Charging per lead, not per account - Intent data should be per-account, not per-lead. If vendor charges $1-5 per lead, you'll overpay.

5. Not transparent on data sources - If vendor won't explain which websites they track, be suspicious.

---

Scoring Models: Rules vs. Machine Learning

Rules-Based Scoring (Abmatic AI, Most ABM Platforms)

How it works: - You define rules: "If Bombora score > 60 AND employee count 100-1000 AND revenue $10M-100M, then intent = 100" - System applies rules consistently - Fast to implement, easy to understand

Advantages: - Transparent (you understand the logic) - Fast to change (update rule, scores update immediately) - No historical data required

Disadvantages: - You have to guess at rule weights - Doesn't learn from results - Can miss patterns in data

Machine Learning Scoring (Metadata.io, 6sense)

How it works: - Feed ML model historical data: "These 100 accounts showed these signals and converted" - ML model identifies patterns (maybe high intent score + recent hiring = 90% conversion) - Model scores all future accounts

Advantages: - Learns from your specific business patterns - More accurate over time - Can identify non-obvious patterns

Disadvantages: - Requires historical data (12+ months) - Black box (hard to understand why an account got a score) - Changes as model retrains

Testing Intent Data: How to Pilot

Run a pilot before committing:

Month 1: Choose vendor and segment

• Pick one intent data source (Bombora, G2, or first-party)
• Identify 100 accounts with highest intent scores
• Commit to 4-week outreach campaign (email + ads)

Month 1-4: Measure engagement

• Track email open rates by intent score bucket
• Track ad click rates by intent score bucket
• Track demo requests from high-intent accounts
• Compare to control group (accounts with no intent signal)

Measure results: - High-intent (70+): 5-8% demo request rate? - Medium-intent (40-70): 2-3% demo request rate? - Low-intent (0-40): 0.5-1% demo request rate?

If high-intent has 5x better response than low-intent, the signal is working.

Month 2-3: Scale if it works

• If intent data is working (high-intent outperforms), expand to 200-300 accounts
• If not working, try different vendor or different signal mix

Cost of pilot: $2.5k-4k for 1-3 months of vendor subscription + internal time

Intent Data for Different GTM Motions

Sales-Led GTM

Use: Bombora (broad market coverage) + Apollo (contact data)

• Identify accounts in-market with Bombora
• Find contacts at those accounts with Apollo
• Sales team outreach to high-intent contacts
• Cost: $3.5k-4.5k/month

Marketing-Led ABM

Use: G2 (research activity) + First-party (engagement confirmation)

• Find accounts researching on G2
• Send targeted ads and email
• Measure engagement on your website
• Convert engaged accounts to sales
• Cost: $2.5k-3.5k/month

Land-and-Expand

Use: First-party (current customer accounts) + Bombora (adjacent opportunity identification)

• Identify expansion accounts using first-party engagement
• Layer Bombora to find accounts showing intent for new products
• Target expansion campaigns
• Cost: $1.5k-2.5k/month

---

Checklist: Evaluating an Intent Data Vendor

Before signing, ask:

1. Data freshness: How often are signals updated? (Daily is good, weekly is acceptable, monthly is too slow)

2. Data sources: Which 30k+ websites are in their consortium? (Ask for list)

3. Accuracy: What's their precision and recall? (Not acceptance rate, actual accuracy)

4. Refresh rate: Do signals decay? (Yes, they should; stale intent is worthless)

5. Coverage: What percentage of target accounts are covered? (Bombora covers ~40% of B2B accounts, G2 ~20%)

6. Integration: Does it sync to Salesforce/HubSpot? (No custom engineering required)

7. Pricing transparency: Is per-account pricing clear? (Avoid "contact sales")

8. Pilot availability: Can you test for 30 days before committing? (Reputable vendors say yes)

Conclusion

Intent data is essential for ABM, but quality varies widely. Bombora offers broad market coverage. G2 offers research activity visibility. First-party data offers highest accuracy for known accounts.

Best approach: Start with one source (Bombora or first-party), pilot for 30 days, measure engagement impact, then expand if working.

Expected accuracy: 60-80% precision for third-party signals, 85-95% for first-party signals. No vendor does better than that.