A Fokals intent score answers one question: on which topics is a company active now, and on what evidence? Each score is one number from 0 to 100 for one company, one topic and one closed week, and it comes with the dated evidence behind it. This post sets out what a row of Intent Scores holds, the 69 topics it covers, what the surge flag marks, and how to read a score and put it to work in a ranking, a model or a call. With it you can open any row and say what it tells you.
What a row of Intent Scores holds
Intent Scores has one row for each company, topic and closed week, and a week runs from Monday to Sunday in UTC.
| Field | What it holds |
|---|---|
| Company | The company's name and its stable company ID, the key to every other Fokals dataset, with market identifiers for a listed company |
| Topic | One of 69 topics, with its name and its group |
| Week | The Monday that starts the closed week the score describes |
| Score | A number from 0 to 100 for the company's activity on the topic in that week |
| Surge flag | Set when a strong score marks a company newly active on the topic |
| Evidence | The dated evidence behind the score |
Each row is written once, after the week closes, and is never revised. A row for a closed week therefore holds what was known when that week ended, which is what a back-test or a model needs. A period written after its normal schedule carries the reconstructed flag, so a strict test can leave it out.
A missing row reads as a quiet week. A company in the index with no row for a topic was quiet on that topic in that week, and the next active week brings it back.
The intent dataset delivers two more datasets beside the scores. Company Signals gives dated company signals daily, each with its kind and the topics it bears on, as events to use in their own right. Company Funding gives private capital raises reported in regulatory filings, and funding announcements. All three join on the company ID.
The 69 topics in nine groups
The topics name what a company may buy, build or expand into. They sit in nine groups.
| Group | Examples of its topics |
|---|---|
| Advertising | Paid social advertising, search advertising, retail media, affiliate and partnerships |
| Marketing | Marketing automation, email and SMS marketing, customer data platform, web analytics |
| Commerce | E-commerce platform, payments and checkout, logistics and shipping, physical retail |
| Bookings and events | Appointment booking, reservations and online ordering, ticketing and events |
| Sales and service | CRM, sales engagement, customer support, live chat and conversational |
| Data and technology | Data warehouse and BI, cloud infrastructure, security, identity and access |
| Website technology | Web frameworks and libraries, web performance, maps and location, web accessibility |
| Business systems | ERP and finance systems, HR and payroll, recruiting, legal and compliance |
| Growth | International expansion, building a new function, new funding |
Choose the topics that describe what you sell or study, and read each one on its own. A score ranks companies within a topic, and two topics are two separate questions: a company can be busy on CRM and quiet on payments in the same week. The data dictionary lists every topic with its group.
The surge flag
The surge flag marks a change from a company's own normal on a topic. A high score says that a company is active. The flag says that it has become active, measured against its own record on the topic, which a high score alone does not show.
Take Acme Robotics, an illustrative company, on two of its topics in the same week.
| Topic | Score this week | Surge | How to read it |
|---|---|---|---|
| CRM | 58 | Yes | Newly active: the account joins this week's CRM list, and the evidence is fresh |
| Web analytics | 74 | No | Steadily active: it has scored high on the topic for months |
Web analytics has the higher score and no flag, because a company that scores high on a topic every week is steadily active on it, and the flag is kept for the newly active. Sort on the score when you want the busiest accounts and on the flag when you want the newly busy ones. Both are written once each week, so a back-test reads either exactly as it stood.
How to read a score
- One topic at a time. Each row is one company, one topic and one week. Rank within a topic, never across topics.
- Read the evidence before the number. The evidence says what stands behind a score and when it happened. Two scores of the same size can rest on very different evidence, and the evidence is what a rep opens a call with.
- Read the level and the change. A score near the top of the scale and a score that has climbed for three weeks say different things. Keep the weekly history of each account and topic, and read the latest row beside the ones before it.
- Read bands, not points. Group scores into bands, treat the accounts inside one band as level, and let the evidence decide which comes first.
- Read the dates. The evidence is dated, so the age of the newest item tells you how recent the activity is on the day you act.
The query below lists the week's CRM scores with their evidence, the newly active accounts first.
select company, company_id, score, surge, evidence
from company_intent_weekly
where topic = 'crm'
and week_start = date '2026-09-21'
order by surge desc, score desc;The guide to prioritising accounts with intent scores joins the score to fit and turns it into a weekly ranking.
What a score gives you
A Fokals intent score is evidence-backed: every score comes with the dated evidence behind it, so a rep, an analyst or a model reviewer can open it and check it. Three uses follow from that.
- A reason to call. The score says where a company's recent activity is concentrated, and the evidence gives the rep a dated fact to open with.
- A feature for a model. Weekly rows are written once after the week closes, and a definition holds still once released, with at least 90 days' notice before a breaking change. A feature built on the score keeps its meaning from one quarter to the next.
- A point-in-time series. Each row is what was known when its week closed, so a back-test reads the score exactly as it stood.
Calibrate the score on your own outcomes. For each band of scores, count the accounts that went on to open an opportunity, and set your cut-off where the bands separate. Keep the cut-off fixed for a quarter so that the bands stay comparable, and hold out a random share of accounts so the rule is tested against what would have happened without it.
Scores arrive weekly and Company Signals daily, by REST API or as bulk files in JSON, JSON Lines or CSV. The methodology describes each dataset, and the comparison of company signals vs third-party intent data sets the outputs of the two kinds of intent data side by side.
Frequently asked questions
What is a Fokals intent score?
One number from 0 to 100 for one company, one topic and one closed week, with the dated evidence behind it. The score says how active the company was on the topic in that week, and the evidence says what stands behind the number. There are 69 topics in nine groups, from advertising and marketing to business systems and growth, and each weekly row is written once, after the week closes.
Why did a company's intent score go down?
Because its activity on the topic went quiet. A score describes a company's recent activity on a topic, so it follows that activity down as well as up. Check the dates on the evidence: when the newest item is several weeks old, the account is going quiet on the topic. A fall records a quiet period, which is a different thing from a change of plan, and fresh activity lifts the score again.
What is a good intent score?
The right cut-off depends on your market, so set it from your own outcomes. Compare companies within one topic, read the evidence, and test bands of scores against your won and lost deals. For a ready-made rule, use the surge flag, which marks a company that has become active on a topic against its own record.
What evidence comes with a Fokals intent score?
Every weekly score comes with the dated evidence behind it. A rep can open what stands behind a number before a call, and a model reviewer can trace a feature back to dated facts. Company Signals, delivered daily in the same dataset, gives the dated company signals as events in their own right, each with its kind and the topics it bears on.
Can intent scores be used in a model or a back-test?
Yes. Each weekly row is written once after the week closes and is never revised, so it holds what was known at that time and a back-test reads it as it stood. A definition holds still once released, and a breaking change ships with at least 90 days' notice, so a feature built on the score keeps its meaning while your model is in use.
How often are Fokals intent scores updated?
Once a week. A score is written for each company and topic after the week closes, Monday to Sunday in UTC, and Company Signals is delivered daily. Both arrive by REST API or as bulk files in JSON, JSON Lines or CSV, so a weekly pull keeps a ranking current and a daily pull keeps the events current.
The queries and code on this page are examples to adapt. Test them in your own environment before you rely on them.