Why intent data scores differ between vendors for the same account

Intent data scores differ between vendors for the same account because each vendor observes different behaviour, scores a different unit, measures against a different baseline and prints the result on a scale with its own meaning. This comparison reads the public documentation of Bombora, 6sense and G2 to set out what each intent data score is built from and what each intent surge is measured against, and sets what Fokals delivers for the same account beside them. With it you can take two numbers for one account, say why they disagree, and run a test that shows which to act on.

Why intent scores for the same account differ between vendors: five causes

Two vendors can give one account different intent scores and both be correct, because each score measures something specific. Five things vary from vendor to vendor: the behaviour that is observed, the unit that is scored, the baseline and window, what the scale means, and the cut-off and refresh that decide what you see on a given day.

1. The behaviour observed. Reading on publisher sites, pages opened on a review site, a customer's own CRM activity and a company's public acts are different events. Ask what is observed, and where.
2. The unit scored. A score can belong to an account and a topic, to an account as one customer's model sees it, or to a buyer in relation to one vendor's product. Ask which.
3. The baseline and the window. Three weeks against twelve and seven days against twenty-one are different questions, and the same activity can answer yes to one and no to the other. Ask which periods are compared.
4. The meaning of the scale. A number out of 100 can express change against a baseline, a modelled likelihood of an opportunity, or the level of recent activity. Ask what the middle of the scale means.
5. The cut-off and the refresh. The line that turns a score into a flag is a setting, and one score can fall on either side of it in two tools. Ask where the line is, who sets it and when the value was computed.

Intent scoring methodology at different B2B intent data vendors: what an account score is built from

The intent scoring methodology of each vendor starts from a different observation. Bombora scores research across its Data Co-op of publisher and brand sites, G2 scores research on G2, and 6sense models each customer's own engagement data with keyword and partner intent.

The Bombora column follows its Company Surge page and its FAQ, the 6sense column its Intent Model and Predictive Modeling Overview articles, and the G2 column its Buyer Intent documentation.

The unit explains many disagreements. 6sense describes a model set up for the customer's organisation, on the customer's own definition of a relevant opportunity, so its score answers one customer's question about an account. G2 sends signals at the level of the vendor's product, so two vendors in one category can hold different evidence on the same buyer. The difference between reading and company signals is set out in company signals vs third-party intent data.

Intent data surge baselines at Bombora, 6sense and G2: the calculation and methodology of each

Bombora, 6sense and G2 document three different measures. Bombora compares an account's latest three weeks of research on a topic with its previous twelve. 6sense measures momentum as the last 7 days against the previous 21, beside a predictive score. G2 reports a stage and an activity level, and we found no baseline in its Buyer Intent documentation.

Bombora Company Surge: the intent data, the baseline and the score

Bombora Company Surge is intent data derived from research across its Data Co-op, and each score compares an account with its own past. Bombora's site describes scoring every pairing of account and topic by setting its latest three weeks of activity against the twelve weeks before them, so the score expresses change and the baseline is the account's own.

Its FAQ on what the score is names four factors: interactions and unique people researching the topic, each counted over the last three weeks against the previous twelve, the weight of the topic within each piece of content, and the depth of engagement, such as dwell time and scroll depth. A second answer adds that activity in the recent three weeks is weighted up to 5% higher than activity in the baseline period.

The FAQ on the score places scores between 40 and 60 at the same or average level of research, lower scores at a decrease and higher scores at an increase, and Bombora's guidance on thresholds treats 60 or more on a topic as spiking. Its FAQ on missing scores says the activity a score requires includes activity on the topic in the first week of the baseline and in at least one of the three most recent weeks. And each geography has its own baseline, so one office location can score differently from the company in the same week.

6sense buying intent: the score, the baseline and the momentum behind a surge

6sense expresses buying intent as a predictive intent score from 0 to 100 and reads five buying stages from it. Its documentation describes the score as the likelihood that an account opens an opportunity in the next 90 days, modelled on the account's historical activity. Change is measured apart, as momentum: the last 7 days against the previous 21.

The score is updated daily, and the buying stages are bands of it: Target from 0 to 19, Awareness from 20 to 49, Consideration from 50 to 69, Decision from 70 to 85 and Purchase from 86 to 100. Baseline has two senses in the 6sense articles we read: the Consideration stage is described as activity more significant than the account's baseline, and in the Model Insights report the baseline of the account behaviour model is a conversion rate, the share of accounts showing any intent in a month that opened an opportunity within three months.

Momentum is defined in the Temperature model: positive when the current week holds more activity than the prior three, calculated on the last 7 days against the previous 21 for each type of activity. 6sense's FAQ on Bombora data defines a surging topic inside 6sense as a Company Surge score of 70 or more in the most recent week, a threshold that is fixed there.

G2 Buyer Intent: two components in place of one score

G2 Buyer Intent reports two components for each buying company, a Buying Stage and an Activity Level, in place of a single score. G2's documentation says both are calculated from the signals in the vendor's subscription and updated daily. The Buyer Intent documentation states no baseline and no look-back window.

Buying Stage is described as an evaluation of the buyer's journey across G2 from factors that include page engagement and signal type, and Activity Level as an aggregate of activity such as time spent per page and pages visited. The documentation adds that Buyer Intent includes research activity from Capterra, Software Advice and GetApp. The vendor's plan determines which signal types it receives, so a stage of decision is a statement about research on those sites in relation to one product.

Surge score methodology side by side: a baseline comparison for B2B intent data

Compared on method, a surge score has four parts: the recent window, the baseline, the line that turns a score into a surge, and the refresh. The documented baselines range from the previous 21 days to the previous twelve weeks, and the documented lines sit at scores of 60 and 70.

The weekly sync is the one Bombora's Salesforce guide documents for that integration, which runs over the weekend.

What Fokals delivers for the same account

Fokals Intent Scores give one score from 0 to 100 for each company, topic and closed week, across 69 topics in nine groups. The surge flag marks a company that has become active on a topic against its own record. Each row comes with the dated evidence behind the score, so the reason for a score can be opened and checked, and rows are written once after the week closes and never revised.

Set beside the three measures above, a Fokals row adds two things to a comparison: a reason a person can open, and a value that stays as it was written, so a test months later reads the row you acted on. What a Fokals intent score gives you shows how to read one, and the methodology describes the dataset.

A worked example: one account, one week, four readings

The example is illustrative: Acme Robotics is an invented company and every figure is an assumption. A plain ratio of weekly averages stands in for each vendor's model, which takes more inputs than a count. The ratio is not any vendor's actual formula or output: it isolates the effect of the window and the cut-off.

Assumptions:

• The topic is CRM software, and week 15 has just closed.
• People at Acme Robotics made 4 research interactions a week on the topic in weeks 1 to 12, then 9 in week 13, 15 in week 14 and 6 in week 15.
• A reading-based score for the account and topic stands at 65 in week 15.
• Its Fokals score on the CRM topic is 58 in week 15, with the surge flag set and two dated items of evidence.

The working:

1. Three weeks against twelve, the windows Bombora documents. The recent weeks hold 9 + 15 + 6 = 30 interactions, or 10 a week. Weeks 1 to 12 hold 48, or 4 a week. Ten against four is 2.5 times the account's own baseline: a rise.
2. Seven days against twenty-one, the windows of 6sense Momentum. Week 15 holds 6 interactions. The prior three weeks hold 4 + 9 + 15 = 28, or 9.3 a week. Six is below that, so the direction is negative: the peak has passed.
3. One score, two cut-offs. A score of 65 is above the 60 that Bombora's guidance treats as spiking and below the 70 that defines a surging Bombora topic inside 6sense.
4. A weekly row with its evidence. The Fokals row for week 15 reads 58 with the surge flag set, and its two dated items of evidence travel with it, so a rep can see what stands behind the flag before acting on it.

The four readings give a rise, a fall, a split and a flag for one account in one week, and each is a correct answer to its own question. If reading returns to 4 a week in weeks 16 to 18, the twelve-week baseline for week 18 is weeks 4 to 15, which hold 36 + 30 = 66 interactions, or 5.5 a week. Four against 5.5 is 0.73, so an account back at its usual level reads below baseline, because the spike now sits inside the baseline.

A test to run before you act on two scores

When two sources disagree on an account, your own outcomes decide which to act on.

1. Take one closed week and a list of your target accounts, matched across sources by domain matching.
2. Turn each source into a yes or no for that week by its own documented rule. Keep the numbers apart.
3. Sort the accounts into four cells: both sources yes, the first alone, the second alone, neither.
4. For each cell, count the accounts that opened an opportunity in the following 90 days, and divide by the size of the cell.

Repeat over several weeks, then rank on the cells with both columns visible, as in prioritising accounts with intent scores.

A look-back needs each value as it stood in its week. 6sense's Model Insights article says its models are monitored and retrained, with at least two weeks' notice of an update, and that accounts can change buying stage on the day an update is rolled out. So store the value you acted on from every source, with its date.

Frequently asked questions

What this page says about the products it names was checked against their public documentation on 7 October 2026. Product and company names are trademarks of their owners. Fokals is not affiliated with them or endorsed by them.