An applicant tracking system, usually shortened to ATS, is software that stores every applicant in one place and moves them through the stages of your hiring process. It holds the job, collects the applications, keeps the notes and the scores, and records who decided what. That is the whole idea. Everything else vendors sell you sits on top of that one job.
If you are hiring from a shared inbox and a spreadsheet, you already have an applicant tracking system. It is just a bad one. The question is not whether to track applicants, it is whether the way you track them is costing you good people.
What an ATS does day to day
Strip away the feature lists and an ATS does five things.
It publishes the job, once, to your careers page and out to the job boards you use, rather than you retyping the same advert into six sites and then trying to remember which one you posted on.
It collects every application into one record per candidate, with the resume attached and the application answers alongside it. No forwarded emails, no attachments lost in a thread, no version of the CV that only one person has.
It moves people through stages you define. Applied, screening, interview, offer. Anyone on the hiring team can see where every candidate sits without asking.
It holds the evidence. Interview notes, scores, who said what and when. This is the part people underrate until the first time somebody asks why a candidate was rejected and nobody can answer.
It tells you what happened. How many people applied, where they came from, how long each stage took, where candidates dropped out.
That is the product. A good ATS does those five things without getting in the way.
The problem it actually solves
The obvious pitch is time saved, and it is real. The bigger cost is the one nobody measures: good candidates lost to disorganisation.
A strong applicant applies on a Tuesday. Nobody looks until the following Monday because the inbox is shared and everyone assumed someone else had it. By then they have two other processes running and yours is the slow one. You never find out you lost them, because the loss shows up as silence.
Stage confusion has the same effect. A candidate has a good first interview, the hiring manager is away, and the thread goes quiet for a fortnight. Nothing went wrong exactly. Nobody decided to reject them. They just fell through a gap that a status field would have closed.
An ATS does not make anyone a better interviewer. What it does is remove the failure modes that have nothing to do with judgement.
Which features are worth paying for
Vendor feature lists are long and mostly undifferentiated. These are the ones that change how the process runs.
One posting to many boards. Sometimes called job distribution. Write the advert once, publish to your careers page and out to the boards. Check whether it is included or billed separately, because that varies a lot between vendors.
Resume parsing that you can correct. Parsing pulls structured fields out of an attached CV. It is useful and it is never perfect. What matters is whether you can fix a mis-parsed record, or whether the bad data is stuck there.
Configurable stages. Your process is not the vendor's default. If you cannot rename stages and add or remove them per role type, you will end up bending your hiring to the software.
Scorecards and structured notes. The single biggest quality lever in hiring, and the one most systems treat as an afterthought. You want the same questions and the same rating scale in front of every interviewer, with the notes attached to the candidate record rather than living in somebody's notebook. See the structured hiring playbook for how to build the scorecard itself.
Interview scheduling that reads calendars. Removes a genuinely tedious loop.
A talent pool you can search later. Roughly a third of your best future candidates are people who applied for something else six months ago. If you cannot search past applicants by skill, you are rebuying attention you already paid for.
Export. Underrated and worth checking before you sign. You should be able to get your candidate data out, in a usable format, whenever you want. If a vendor is vague about this, that is the answer.
Features that sound impressive and rarely change an outcome: sentiment analysis on candidate messages, gamified assessments bolted into the pipeline, and any scoring that ranks people without showing you why.
The entry-tier trap
This is the part to read twice, because it is where most buyers get caught.
Many vendors price by advertised tiers, then move the features you actually need up a level. Scorecards on the middle plan. Reporting on the plan above that. Single sign-on treated as an enterprise upsell. The entry price gets you in and the real price arrives at renewal, once your data is in and moving is painful.
Before you compare prices, write down the six things you will genuinely use, then check which plan each one is on. Compare the plans that contain your six, not the headline numbers. The cheapest entry tier is frequently the most expensive system over three years.
Two specific things worth checking:
Per-user pricing. If every hiring manager needs a seat and seats are billed individually, the system gets more expensive exactly as you get better at involving the hiring team. That is a bad incentive to build into your process.
Security as an upsell. Single sign-on gated behind the top tier is common and hard to defend. Security is not a premium feature.
What an ATS costs
Pricing falls into three shapes.
Per job. You pay for each role you run. Suits low, irregular hiring volume. HRM8 prices this at US$195 per job.
Per month by volume. You pay for a band of concurrent open roles. This is the common shape once hiring is steady. HRM8's plans run from US$295 per month for up to five open roles, US$495 for up to ten, US$695 for up to twenty, and from US$995 for unlimited, billed annually. Every paid plan includes unlimited users.
Per employee. Less common for standalone applicant tracking, more common when the ATS is bundled into a wider HR platform.
There is also a free tier worth knowing about, because it changes the calculation for small employers. HRM8's Lite plan is free and covers three users and one open job, with job distribution included, which is enough to run a real hire properly rather than out of an inbox. Full detail sits on the pricing page.
As a rule of thumb: if you are filling fewer than about five roles a year, pay per job. Beyond that, a subscription is usually cheaper and always less friction.
Do you actually need one yet
An honest set of triggers, rather than a sales answer.
You probably need one if more than one person is involved in hiring decisions, if you run more than one role at a time, if candidates regularly wait more than a week for a response, if you cannot answer "how many people applied for that role and where did they come from", or if you would struggle to show why a specific candidate was rejected.
You probably do not need one yet if you hire once or twice a year, one person owns the whole process end to end, and you can hold every candidate in your head. A shared spreadsheet and a disciplined process will serve you, and the discipline matters more than the tool.
The trigger is rarely headcount. It is the number of people involved in the decision. One person with a spreadsheet is fine. Four people with a spreadsheet is a mess.
Getting one live without the usual pain
Implementation failures follow a pattern, and it is almost never the software.
Do not migrate everything. Bring across active candidates and the talent pool you will genuinely search. Historical applicants from four years ago are not an asset, they are a data-protection obligation.
Set up one role first. Configure your most common hire, run a real vacancy through it end to end, then build the rest from what you learned. Configuring all twelve role types before anyone has used the system produces twelve wrong configurations.
Decide who owns it. One person, named, responsible for the configuration. Systems without an owner drift back to email inside a quarter.
Bring the hiring managers in early. They are the users who will quietly go around the system if it feels like admin imposed on them. Show them the scorecard view and the fact that they no longer have to chase recruiters for status, and they will stay in it.
The realistic window for a small team is days, not months. Anyone quoting you a long implementation for straightforward applicant tracking is describing their own complexity, not yours.
Where AI genuinely helps, and where it does not
AI in hiring is oversold in both directions. Two distinctions hold up.
It is genuinely good at ordering and summarising: surfacing which applications match the criteria you defined, pulling a long list into a sensible reading order, drafting the first version of a job description, summarising an interview into structured notes.
It is not reliable at deciding. Ranking people on inferred qualities, scoring "fit" from a video, or auto-rejecting anyone should be treated as out of bounds. Those systems learn from past decisions, and past decisions carry past bias forward at scale.
The line worth holding is simple: AI narrows and drafts, humans decide. HRM8's position is that nothing is auto-rejected, and the employer owns the hiring decision. If a vendor cannot tell you why a candidate was scored the way they were, you cannot defend that decision, and increasingly you may be required to.
For the detail, see where to automate hiring.
Common questions
What does ATS stand for?
Applicant tracking system. It is software that stores applicants in one place and moves them through the stages of a hiring process.
Do small businesses need an applicant tracking system?
Not always. The trigger is how many people are involved in the decision rather than company size. One person hiring once a year does not need one. Three people running two roles at once will lose candidates without one. Free tiers exist, including HRM8's Lite plan, so the cost barrier is lower than it used to be.
How much does an applicant tracking system cost?
Either per job, from around US$195, or per month by open-role volume, commonly US$295 to US$995 and up. Watch which features sit on which tier, because the headline price and the usable price are often different plans.
Does an ATS reject candidates automatically?
It should not, and a well-designed one does not. Filtering and ordering is reasonable. Automatic rejection is not, and in several jurisdictions automated decisions on employment now carry specific obligations. HRM8 never auto-rejects a candidate.
Can a recruitment agency use an ATS built for direct employers?
Not in HRM8's case. The platform is direct-employer-only. Agencies may use it for hiring their own staff, but cannot resell it to service external clients.
An ATS is a database with a workflow on top, and its real value is removing the ways good candidates get lost through disorganisation rather than through judgement. Choose on the six features you will genuinely use, check which plan each of those sits on before you compare prices, confirm you can export your data, and configure one role properly before you build the rest.