Buyer guide · MinMaxHR
The MinMaxHR buyer's guide
Application volume is not the problem most hiring teams think they have. Recruiters have a fixed amount of attention per role, and a stack of resumes spends it in the order the files happened to arrive. The question worth answering is not how to get through all of them. It is who deserves attention first, and why.
This guide is the short version of how MinMaxHR answers that: what the system does at each step, what it will not do, what it costs, and how to check the ranking against a role you have already closed before you spend anything.
When the resumes keep coming, MinMaxHR helps your team decide who deserves attention first, why they fit, what is missing, and where a human should review.
Why is resume screening still the bottleneck?
Because reading is not the same as evaluating. A single corporate opening commonly draws 200 to 250 applications, and eye-tracking research by The Ladders found recruiters spend about 7.4 seconds on an initial resume scan. Seven seconds is a glance. The constraint is not reading speed. It is that attention runs out before the pile does.
The old question was how do we get through all these resumes, and it makes reading speed the goal. The better question is who deserves our attention first, and why. That one makes decision quality the goal, and speed follows from it rather than the other way round.
The cost of getting this wrong is not a missed deadline. It is a strong candidate sitting at position 180 in a queue nobody reached, while the role stays open and the team keeps interviewing whoever surfaced early.
How does MinMaxHR actually work?
Six steps, and each one exists because the step before it changes what your team can trust.
- A role opens. The job description becomes the thing every candidate is measured against, so fit means one specific role rather than a general impression.
- Applications arrive. Every file is identified and read, including scanned documents and photographs, and duplicates are caught before any scoring runs.
- Candidates need prioritising. CandidRanker scores each one from 0 to 100 against that job description, not against a keyword list.
- A ranking appears. Matches, gaps and every contributing factor are inspectable, including the ones that scored zero.
- Recruiters review. Mandatory and optional requirements, plus your own terminology, shape how the evaluation runs.
- People decide. Shortlist, select and reject are human actions. Each needs a written reason and is recorded against a named recruiter.
CandidRanker ranks and explains. Your people decide. MinMaxHR records the evidence. If you remember one sentence from this guide, that is the one.
What is the difference between input quality and candidate fit?
They are two different questions and collapsing them is the most common way screening tools go wrong. Input quality asks whether a document contains enough reliable information to evaluate at all. Candidate fit asks how well the person matches the role.
A strong candidate can submit a badly formatted resume. A beautifully formatted resume can describe a poor fit. Keeping the two apart is what stops a tool quietly rewarding formatting over substance, and it tells the recruiter which problem they are actually looking at. The methodology page sets out how the scoring itself works.
Is this the same as pasting a resume into ChatGPT?
No, and the difference is worth understanding before anyone on your team does it. Pasting a candidate's resume into a public chatbot sends their data to a system you do not control, under terms you did not review, with no record of who saw what. Your organisation stays accountable for that data either way.
CandidRanker's AI processing and its database are hosted in India. Where an AI assistant is involved, it connects through the MCP server: authenticated by OAuth, scoped to one workspace, carrying only the permissions you approve, and writing every action it takes to the same decision log as a recruiter's. Same assistant, entirely different control surface.
Under the Digital Personal Data Protection Act your organisation is the Data Fiduciary for candidate data and a screening vendor acts as a processor on your instructions. This guide is general information about a real consideration, not legal advice, and the DPDP page goes through it properly. Controls and data handling are documented on the security page.
Who is MinMaxHR built for?
Five kinds of team buy this, and each one is protecting against something different.
- Talent acquisition. Wants faster, more consistent shortlists. Worried that a tool will quietly replace recruiter judgement. It improves the first decision and leaves the decision itself with people.
- Staffing and recruitment agencies. Want repeatable, client-ready shortlists. Worried that every client needs different criteria. Criteria, terminology and evidence are set per workspace, one per client.
- High-volume hiring. Wants attention going where it matters first. Worried screening stays the bottleneck regardless. Volume becomes a prioritised review queue instead of a pile.
- Technical and manufacturing. Wants complex terminology handled properly. Worried that generic matching reads straight past the context. Your plant's vocabulary and aliases are defined once and applied consistently.
- HR leadership. Wants quality, consistency and governance. Worried that AI becomes something nobody can account for. Every human decision stays explicit, attributed and inspectable.
What does CandidRanker cost?
The MinMaxHR platform is free. Accounts, workspaces and team invitations cost nothing. CandidRanker is priced by volume: Free covers 100 resumes and 10 job descriptions at no cost and no card, and Growth is ₹9,999 per 30 days for 1,000 resumes and 100 job descriptions. Enterprise is unlimited volume with REST API and MCP integration.
Paid plans are a one-time payment that activates a 30-day period. No auto-renewal, no lock-in, nothing to cancel. A plan covers one account and your colleagues can be invited into the same workspace and share its limits, so the price is per workspace rather than per seat. Monthly and annual invoiced billing is available for teams that need it. Full pricing is here.
How do I evaluate this without sitting through a demo?
Run a retrospective test. Take a role you have already closed, upload the job description and the resumes you actually received, and compare the ranking against the people your team shortlisted and hired. It takes about an hour, it runs on the free plan, and it uses your data rather than a curated demo set.
It is the fastest honest way to find out whether this works on your hiring, because you already know the right answer. The retrospective test is described step by step here.
If you would rather talk it through first, or you need invoiced billing, volume pricing or a security review, write to sales@minmaxhr.com and a person will reply. You can also download this guide as a PDF to share internally.
Frequently asked questions
- What is a hiring decision system?
- It is the layer that decides who your team looks at first and records why. Your applicant tracking system owns the pipeline: postings, stages, candidate records. A hiring decision system owns the decision step, ranking every applicant against one specific job description, showing the evidence, and logging who decided what.
- Does MinMaxHR replace our ATS?
- No. It sits next to it. Keep Workday, Greenhouse, Lever, Zoho Recruit or whatever your team already runs. The ATS keeps managing the workflow while MinMaxHR handles evaluation, ranking and the decision record.
- Will it reject candidates automatically?
- No. Shortlist, select and reject are human actions. Each one needs a written reason and is logged to a named recruiter. There is no auto-rejection setting to turn on and no pending state the system moves candidates into on its own.
- What if it ranks someone wrong?
- You inspect the evidence behind the ranking and decide differently. Every dimension shows its reasoning, including the ones that scored zero, so a result you disagree with is something you can open up rather than argue with. Scoring is deterministic, so the same job description and the same resume produce the same score every run, which is what makes a retrospective test meaningful.
- Where is candidate data stored and where does the AI run?
- Customer data is stored and processed in Mumbai, India, on Google Cloud asia-south1, in tenant-isolated workspaces with TLS in transit and encryption at rest. CandidRanker's AI processing and its database are hosted in India.
- Is the price per user?
- No. A plan covers one account and you can invite your team into the same workspace, where they share that plan's resume, job description and decision limits. The price is per workspace, not per seat, so adding a colleague does not add a fee.
- How do I try it on my own data before paying?
- Use the free plan on a role you have already filled. Upload that job description and the resumes you actually received, then compare the ranking against who your team shortlisted and hired. It takes about an hour and needs no demo and no card.