TalentNeuron review: what it sells, what the API meters, and who it is built for

TalentNeuron sells labour market intelligence to enterprise HR. Not job postings and not a feed: dashboards and forecasts that tell a workforce planning lead where the talent for a role is, what it costs there, who else is competing for it, and what the organisation will need three years out. The site says the platform is “trusted by 60% of the Fortune 100”, and the buyer it is written for sits on a committee rather than at a terminal.

TalentNeuron does sell an API. It is metered in units it does not publicly define, gated by your subscription tier, and scoped to your own internal business purposes. If you are building a product on hiring data rather than answering a planning question inside one company, that is a different shape from what you need. Here is what the platform actually is, where it came from, what the API terms commit to in writing, and how to reproduce the two questions most teams buy it for.

Where TalentNeuron came from

The lineage matters, because it explains the product. Per the company’s own about page, one half started as Wanted Analytics, founded in Quebec City in 1999; the other was founded as Talent Neuron in Bangalore in 2011. CEB acquired and merged them around 2014 to 2015, and Gartner acquired CEB in 2017. TalentNeuron then spent six years inside a research and advisory giant, which is where its instincts were formed.

In February 2023 Gartner sold it to Leeds Equity Partners. Gartner’s own filing describes the disposal of a non-core business for approximately $161.1 million after post-close adjustments. In 2024 TalentNeuron acquired HRForecast, a German workforce analytics company. So the thing you are buying today is a twenty-five-year-old data asset that has been an analytics vendor, a consultancy annex, and now a private-equity-backed independent, in that order.

What you actually buy

The catalogue splits four ways: Strategic Workforce Planning, Workforce Intelligence, Data as a Service, and Strategic Consulting. Inside the platform sit the modules that do the work, and their names give away the use cases better than any brochure copy: Hiring Analysis, Competitor Analysis, Skills Analysis, Location Analysis, Employee Value Proposition tracking, Organizational Design, Skills and Career Marketplace, Work Architecture, and an assistant called Synapse.

Read that list as a shopping list of questions. Where should we open the next office. What are our three competitors paying for this role in Krakow. Which skills are drifting out of our job family. Those are real questions and they are worth money. They are also, critically, questions about aggregates rather than about individual postings.

There is no published price. Every route through the site ends at “Book a Demo”, and Data as a Service is sold the same way. If you have priced this category before, the shape will be familiar from Lightcast pricing: an annual contract scoped by seats, markets and modules, quoted after a discovery call.

The API, read line by line

This is the part worth doing carefully, because the public terms page for the Talent Intelligence API is more specific than the marketing pages are. Eight API families are named: Hiring Indicators, Location Strategy, Skill Strategy, Competitive Benchmarking, Diversity & Fresh Talent, EVP Insight, Taxonomy Foundations and Taxonomap. Which of them you can call is “determined based on client’s base membership”, so the API is an extension of a platform subscription rather than a product you can buy on its own.

The meter is a “Call Unit”. The terms set a rate limit of “120 Call Units per minute per level” and a service limit of a “Quota of 1,000,000 Call Units”, and they reserve the right “to suspend the account after the call unit quota is exhausted”, with reasonable endeavours to return a descriptive error first. Suspension rather than throttling is a meaningful choice to know about before you put the API behind anything that has to stay up.

What a Call Unit costs per request is not on the terms page. It links out to a support article on Call Units and API levels, and that link now returns a 301 to the support portal home rather than the article. So the one number you would need to model spend is, as of this week, not publicly readable. Ask for it in writing during the demo, per endpoint, before you size anything.

The scope clause is the other thing to read closely. The service “enables the Client to access and utilize human resources data for their internal business purposes”, with the full Data Services Terms governing the rest. Internal-use language is normal for enterprise data contracts and it is exactly the clause that decides whether you can put the output in front of your own customers. If you are building a product, get redistribution addressed explicitly rather than inferred.

Read the coverage numbers twice

TalentNeuron publishes a lot of scale: 200+ unique markets, 65,000+ skills, 3 billion profiles, 40+ terabytes of normalized data, 26+ million companies tracked, more than two decades of history, and data representing “more than 96% of the global GDP”.

The daily-volume claim is worth a second look, because the unit changes between pages. The homepage says “3 million job postings in 33 languages processed daily”. The technology page says the platform “captures more than 7 million job posting data points from 142 countries” each day, across 5,000+ metropolitan areas and more than 40,000 sources. Postings and posting data points are not the same unit, and a data point can be a field on a posting. Neither claim is wrong; they just are not the same measurement, and only one of them is a volume you could compare to another vendor.

That is the general problem with this category, and the reason we keep writing about it: coverage claims are stated in units chosen by the seller. The useful version of a coverage claim is one you can reproduce.

Doing the location question yourself

Location Analysis is the flagship use case, so it is a fair test. The question is: for one occupation, where is the volume and what does it pay. On JobsPipe every live posting carries an ISCO-08 occupation code, so the query is one call. Software developers are ISCO unit group 2512, one of the 433 unit groups in ISCO-08.

curl -s -X POST https://api.jobspipe.dev/v1/jobs/search \
  -H "Authorization: Bearer $JOBSPIPE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "occupation_code_or": ["2512"],
    "job_country_code_or": ["US", "IN", "CA", "DE", "SG"],
    "posted_at_max_age_days": 30,
    "include_total_results": true,
    "limit": 1
  }'

For the cost half, the compensation endpoint aggregates the advertised bands for that occupation directly:

curl -s -H "Authorization: Bearer $JOBSPIPE_API_KEY" \
  "https://api.jobspipe.dev/v1/insights/occupations/2512/compensation?country=US&window_days=30"

Run the same shape against our live index today and it returns this. Every figure below is a count of active postings, measured this week:

The column every vendor should show you

The third column is the point. India has three times the software-developer posting volume of Canada and roughly one fifty-fifth of the salary evidence, because employers there mostly do not publish pay in the ad. Any platform can render a median for India from those 67 rows. It will look exactly as authoritative as the American one, drawn from 46,601.

Across the whole live index, 29.3% of active postings carry a salary at all, and that share is wildly uneven by country. Meanwhile 86.5% carry an ISCO-08 occupation code, 88.5% carry at least one normalized skill, and 73.5% carry an ISIC industry division. Demand and skills signal is dense. Advertised pay is not, anywhere, for anyone. A vendor that shows you a salary chart without the sample size behind it is not showing you more data than we have; it is showing you the same data with the caveat cropped out.

Who should buy which

TalentNeuron is a good fit if you are an enterprise HR function, you want finished answers rather than rows, you need the consulting layer and the org-design modules alongside the data, and the output stays inside your own walls. The Fortune 100 penetration is not marketing noise; that is a real installed base solving real planning problems, and reproducing Organizational Design or an EVP tracker from raw postings is not a weekend.

It is the wrong shape if you are a developer. There is no self-serve signup, no public price, no public per-call meter, the API surface is gated by your platform tier, and the licence points at internal use. For building a recruiting product, a sourcing tool or an agent, what you want is the postings themselves with the classification already applied, priced per job and readable in one call. JobsPipe reads 30+ ATS and job-board sources into one normalized schema with ISCO-08 occupations, ISIC divisions, normalized skills and salary on every row, and the free plan is 1,000 jobs a month with no sales call.

For the wider map of who sells what in this category, see labour market intelligence and talent intelligence platforms. This review is based on TalentNeuron’s public site, its published API terms and Gartner’s filings, read this week. Pricing and terms are not published in full, so confirm anything contractual with the vendor.