Intent data is useful to investors researching demand for a software category when it is read across many companies at once. The measure is the share of a fixed cohort of companies that score on the category's topic in each week. This guide builds that share from intent data: the cohort, the query, a worked example, the tests to run and how to set the result beside the figures a vendor reports.
Is intent data useful for investors researching demand for a software category?
Yes, for one question: how many companies are acting on a category, and whether that number is growing. Company-level intent data dates the signals that surround a software purchase, so the share of a fixed cohort scoring on a topic gives a weekly reading of category demand between the dates on which vendors report.
A single score ranks one company for a sales team. An investor needs the other view: one topic, many companies, through time. That view asks three things of the data: the unit is the company, every observation is dated and kept as written, and the topic and the score keep one definition.
The Fokals intent dataset is built on all three. Intent Scores carries a weekly score from 0 to 100 for each company and topic, across 69 topics in 9 groups, with a surge flag when a score rises sharply against the company's own normal and the dated evidence behind every score. Company Signals delivers every dated signal behind the scores, with the topics it concerns. Weekly rows are written once, after the week closes, and are never revised.
| Question about the category | Measure | Dataset |
|---|---|---|
| Is demand broadening? | Share of the cohort scoring on the topic, week by week | Intent Scores |
| Is it new demand? | Companies entering the scoring set during the window | Intent Scores |
| How deep does it run? | Dated signals per scoring company | Company Signals |
| Is it the category or the whole market? | The topic's share among companies scoring on any topic | Intent Scores |
| Where is it concentrated? | Intent surge series by industry and country | Market Series |
Which B2B software buyer intent signals show category demand to investors?
B2B buyer intent signals that carry demand for a software category include research on publisher sites, research on a software marketplace, and weekly company-level scores by topic. Investors can read category demand from each of them once the signals are counted by company and by topic over a base that stays fixed.
Each kind answers a different question, and each provider documents what it delivers.
| Kind of intent data | What it delivers | As documented |
|---|---|---|
| Research across publisher sites | A company's research activity on a topic | Bombora describes Company Surge as derived from its Data Co-op of B2B publishers and brand sites, and scores each account and topic by comparing the most recent three weeks of activity with a 12-week historical baseline |
| Research on a software marketplace | Views of a product's pages and of its category | G2's documentation of Buyer Intent lists signals for views of a product profile, pricing, comparison, alternatives and category pages, sent at the product level |
| Company-level scores by topic | A weekly score from 0 to 100 for each company and topic | Fokals Intent Scores: dated, written once and never revised, with the evidence behind every score |
The comparison of company signals and third-party intent data sets the outputs side by side.
The aggregation below applies to any source that is company-level, dated and fixed in definition.
B2B intent data as a market demand signal for venture capital and investment research
B2B intent data becomes a market demand signal when it is aggregated: count the companies scoring on a topic, divide by a cohort that stays fixed, and follow the share week by week. A venture capital investor reads it to test a market before a round, and public-market investment research reads it between reporting dates.
- Venture capital. Before a round, the question is whether the buyer base of the category is widening in the segment the company sells to, and whether the growth is new buyers. The company index includes verified private companies, so the cohort can be private buyers. The guide to sourcing private companies reads the same data company by company.
- Public equity research. Between results, the question is whether category demand runs ahead of or behind the growth a listed vendor has guided to. Rows for listed companies carry the ticker, MIC, ISIN, LEI and share-class FIGI, so a cohort can be a list of securities.
Intent data licensed for research falls within what the SEC's Division of Examinations calls alternative data: information used in financial analysis beyond traditional financial statements, company filings and press releases. Its risk alert of 26 April 2022 recalls that Section 204A of the Investment Advisers Act requires all investment advisers, registered and unregistered, to establish, maintain and enforce written policies and procedures reasonably designed to prevent the misuse of material non-public information. The staff reported advisers whose diligence of alternative data providers was ad hoc, and advisers with no system for deciding when to repeat it. Every Fokals dataset is collected from public information, and the sourcing statement sets out the collection for a due diligence questionnaire.
The measure: the share of a fixed cohort scoring on a topic
The measure has three parts, and each is a choice to write down before you look at a result.
| Part | Definition | What to fix |
|---|---|---|
| Cohort | The companies in the base, chosen on a base date | The list, for the whole window |
| Scoring company | A cohort company whose score on the topic is at or above your cut-off in the week | The topic, the cut-off and the intent version |
| Share | Scoring companies divided by the cohort size | The base, every week |
A company index grows every day, so a raw count of scoring companies rises with coverage. Fix the base in one of two ways: freeze the list of companies you track on the base date, or take every company observed before the window opens and still observed when it closes. The second is a same-store cohort, the base the weekly Market Series use.
The query builds the cohort from Technology Stack and the weekly counts from Intent Scores, for the CRM topic.
-- Postgres. :first_week and :last_week are Mondays; :cutoff is your score cut-off.
with cohort as (
-- companies observed before the window opens and still observed in its last week
select company_id
from company_technologies
group by company_id
having min(first_seen_at) < :first_week
and max(last_seen_at) >= :last_week
),
weekly as (
select i.week_start,
count(distinct i.company_id) filter (where i.topic = 'crm') as crm_companies,
count(distinct i.company_id) as active_companies
from company_intent_weekly i
join cohort c on c.company_id = i.company_id
where i.week_start between :first_week and :last_week
and i.intent_version = :intent_version
and i.score >= :cutoff
and not i.reconstructed
group by i.week_start
)
select w.week_start,
w.crm_companies,
n.cohort_size,
round(100.0 * w.crm_companies / n.cohort_size, 2) as share_of_cohort_pct,
round(100.0 * w.crm_companies / w.active_companies, 2) as share_of_active_pct
from weekly w
cross join (select count(*) as cohort_size from cohort) n
order by w.week_start;A company enters the cohort when Technology Stack (company_technologies) shows a technology on its website first seen before the first week and one last seen in or after the last week. Intent Scores (company_intent_weekly) holds one row per company, topic and closed week. The filter on intent_version keeps one definition across the window, and the filter on reconstructed leaves out any week written after its normal schedule. A cohort company with no row on the topic had no score written for that week: it stays in the base.
The second count, active_companies, is the control: cohort companies at or above the cut-off on any topic. A company active on many fronts can score on several topics in the same week, so the category is read against all of them.
To see how deep the demand runs, count the signals behind it. Company Signals (company_signals) carries every dated signal behind the scores with the topics it concerns.
-- with the cohort defined as above; topics loaded as jsonb
select cast(date_trunc('week', s.observed_at) as date) as week_start,
count(*) as signals,
count(distinct s.company_id) as companies
from company_signals s
join cohort c on c.company_id = s.company_id
where s.topics ? 'crm'
and s.observed_at >= :first_week
and s.observed_at < cast(:last_week as date) + 7
group by 1
order by 1;Signals per company that rise while the share rises say the demand is deepening at each company as well as broadening across the cohort.
For a cut by industry or country, start from Market Series (market_series), which carries an intent surge family with the topic as its subject, on same-store cohorts over 7, 30 and 180 days. Each row states its count, cohort, rate and growth.
select as_of, dimension, count, cohort, rate, previous, growth
from market_series
where metric = 'intent_surge:crm'
and dimension_kind = 'industry'
and window_days = 30
order by as_of, dimension;An illustrative example: demand for CRM software in a fixed cohort
Acme Robotics is an invented company that sells CRM software to mid-sized manufacturers, and a venture investor is weighing its next round. Every figure below is illustrative.
Assumptions:
- The cohort is 1,200 companies in the segment Acme Robotics sells to, fixed on the base date.
- The topic is CRM, and a company is scoring in a week when its score is 50 or more.
- The window is 13 closed weeks at one intent version, with no late rows. Weeks 1, 5, 9 and 13 are shown.
| Week | Scoring on CRM | Share of cohort | Scoring on any topic | CRM share of active |
|---|---|---|---|---|
| 1 | 84 | 7.0% | 300 | 28.0% |
| 5 | 96 | 8.0% | 312 | 30.8% |
| 9 | 111 | 9.25% | 324 | 34.3% |
| 13 | 126 | 10.5% | 336 | 37.5% |
- Breadth. The share of the cohort is 84 / 1,200 = 7.0% in week 1 and 126 / 1,200 = 10.5% in week 13, a rise of 3.5 percentage points. Scoring companies grew from 84 to 126, which is 50%.
- Control. Companies scoring on any topic grew from 300 to 336, which is 12%. The category grew about four times as fast as general activity, and its share of active companies rose from 84 / 300 = 28.0% to 126 / 336 = 37.5%.
- New and continuing. Of the 126 companies scoring in week 13, 60 were also scoring in week 1 and 66 were not. Of the original 84, 24 went quiet. Entrants are 66 / 1,200 = 5.5% of the cohort, and 66 less 24 gives the net change of 42.
- Adoption. Among the 126, Technology Changes records a CRM tool added to the website within the window for 70 companies: 56% have a new tool in place, and the other 44% are acting on the category without one yet.
The reading: demand in this segment is broadening, the growth is specific to the category, and 66 of the 126 scoring companies are new to the set. The table leaves open how many of the 66 chose Acme Robotics.
What the share says about a vendor's revenue
The share measures the breadth of demand for a category: how many companies in a fixed set are acting on it. A vendor's revenue is that breadth carried through three further steps, each read from its own source.
| Step from category demand to revenue | Where to read it |
|---|---|
| How many companies are acting on the category | The share of the cohort, from Intent Scores |
| Which vendor they chose | Adoptions by technology, from Technology Changes |
| What each purchase is worth | The vendor's reporting: price, plan mix, contract value |
| When it reaches revenue | The vendor's results and its recognition policy |
The second step is measurable for a vendor whose product shows on its customers' websites: count its adoptions against all adoptions in its technology category, as in reading technology adoption as a signal of vendor revenue.
For the last two, read the vendor's definitions before you compare. The SEC's guidance on key performance indicators and metrics, effective 25 February 2020, says the Commission would generally expect a metric in management's discussion and analysis to come with a clear definition and how it is calculated, among other disclosures, and that a company which changes its method should consider disclosing, where material, how the calculation differs from prior periods. Total customers, active customers and net customer additions are among the examples it names. A vendor that counts only customers above a contract size describes a narrower set than your cohort.
- Test the lag. Set the share against reported net customer additions at a lag of zero, one and two quarters, and keep the lag that holds on quarters you did not fit.
- Compare weeks a quarter apart. Neighbouring weekly readings move together, so test the series on weeks 13 apart, which are the more independent readings.
Official benchmarks to set the share against
Two statistical offices publish software adoption as a share of businesses, and both tell you whether your share is plausible.
Eurostat's statistics on cloud computing in enterprises report that 52.74% of EU enterprises used paid cloud computing services in 2025, and that among enterprises buying cloud services 27.91% used them for software that manages customer information. The source is the 2025 EU survey on ICT usage and e-commerce in enterprises, which covered enterprises with at least 10 employees or self-employed persons.
The US Census Bureau's Business Trends and Outlook Survey collects data every two weeks on employer businesses. One of its questions is whether the business used AI in the past two weeks, and the Bureau reported overall AI use at between 17% and 20% of businesses from December 2025 to May 2026. The same report notes that in November 2025 the question was revised from AI used in producing goods or services to AI used in any business function.
Two points carry over to your series.
- Level and flow. A survey counts businesses that use the software. An intent share counts companies scoring on the topic in a week. Compare direction and the ranking of segments, and expect a gap in level.
- Definition. A reworded question changes what the level means. Hold the intent version and the cut-off fixed, and start a new series when either changes.
Tests to run before you rely on the series
- Freeze the cohort. Store the company IDs with the base date. Run the query again a month later: every past week returns the same counts.
- Hold one definition. One intent version, one topic, one cut-off. Then repeat the series at a second cut-off: the direction should agree.
- Count the late rows. Run the series with and without rows flagged as written after their normal schedule.
- Control for general activity. Report the topic's share among companies scoring on any topic, and run the query on a topic outside your thesis. If both rise together, the movement is general.
- Open the evidence. Take twenty scoring companies from the latest week and read the evidence on each row. Record how many read as a company acting on the category.
- Test out of sample. Fix the lag and the cut-off on one set of quarters, then score them on quarters you held back.
The data dictionary defines every field in the queries, and the guide to point-in-time data for back-tests covers how to store each load.
Frequently asked questions
Can intent data predict a software vendor's revenue?
Intent data measures how many companies are acting on a software category, which is one input to a vendor's revenue. The others are the vendor's part of those purchases, the value of each contract and the timing of recognition, and they are read from the vendor's own reporting. Test the series against reported net customer additions at several lags, and keep it as a signal for the vendors where the relationship holds on quarters you held back.
What is a fixed cohort, and why does a demand series need one?
A fixed cohort is a set of companies chosen on a base date and kept unchanged through the window. A company index grows every day, so a raw count of companies scoring on a topic rises with coverage. Dividing by a cohort that was already observed before the window opened, and is still observed when it closes, leaves only the change in activity. The Fokals weekly Market Series are computed this way, on same-store cohorts.
How often are intent scores updated, and are past weeks revised?
Fokals Intent Scores are written weekly, one row per company and topic for each closed week, Monday to Sunday in UTC, and the signals behind them are delivered daily. Each week is written once after it closes and is never revised, so a series built today returns the same past values next year. That property lets a back-test read each score as it stood.
Is intent data alternative data, and what should a fund's compliance review cover?
Intent data used in investment research falls within what the SEC's Division of Examinations described as alternative data in its risk alert of 26 April 2022: information used in financial analysis beyond traditional financial statements, company filings and press releases. The alert recalls that Section 204A of the Investment Advisers Act requires advisers to establish, maintain and enforce written policies and procedures reasonably designed to prevent the misuse of material non-public information. The staff's observations concern the diligence of data providers: whether it covers every source, the terms of collection and when it is repeated. Every Fokals dataset is collected from public information, and the sourcing statement sets out the collection.
Which intent topics correspond to a software category?
Fokals scores 69 topics in 9 groups: advertising, marketing, commerce, bookings and events, sales and service, data and technology, website technology, business systems, and growth. Pick the topic that names the category you study, such as CRM, and add a neighbouring topic only when the vendor sells into both. Keep the choice fixed for the whole window, because a share built on a changing set of topics measures the change in the set.
How do venture capital investors use intent data before a round?
A venture capital investor uses company-level intent data to test the market a company sells into: whether the number of companies acting on the category is growing in the industry and country the company targets, and whether the growth comes from new buyers. Read company by company, the same data also shows which private companies are themselves adopting tools, hiring and raising capital, which supports sourcing.
The queries and code on this page are examples to adapt. Test them in your own environment before you rely on them.
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.