You started a company to build something. Now you're reading about professional tax slabs at 11pm because payroll took your whole Saturday and someone asked about their leave balance for the third time this week.
Looking for HR software?
Check out Techimply's List of the Best HR Software in India for your business.
Most founders buying HR software have no HR background. That's normal, and it creates a specific problem: you're evaluating something where you can't tell the difference between a feature that matters and one that sounds impressive. Vendors know this.
This is a checklist for that situation. It assumes you don't know the jargon, don't have an HR person, and need to make a decision without becoming an expert first.
The Reframe That Makes This Easier
Founders instinctively evaluate HR software the way they evaluate other tools: compare features and pick the richest one.
That instinct is wrong here, and understanding why saves you from most bad decisions.
HR software has one job that is legally consequential and several that are merely useful. The consequential one is statutory compliance: computing PF, ESI, TDS, and professional tax correctly and generating the files you file with the government. Get that wrong and you owe money with interest, and the liability is often retrospective.
The useful ones, leave tracking, performance modules, engagement surveys, and org charts, are conveniences. Valuable, but a mistake there costs you annoyance, not penalties.
So the evaluation isn't "which has more features." It's "which one gets the legally consequential part right and is simple enough that I can actually run it." "A platform with 200 features that computes PF on the wrong base is worse than one with 20 that gets it right.
What You Actually Need to Know About Compliance
You don't need to become an expert. You need to know enough to ask real questions.
The thresholds that trigger obligations
The Provident Fund becomes mandatory at 20 or more employees, with a ₹15,000 monthly ceiling on basic plus DA for mandatory coverage. Below 20, registration is voluntary.
ESI starts earlier, at 10 or more employees in most states, covering employees earning up to ₹21,000 a month.
Two things that catch founders out:
The count isn't just your full-time staff. It includes contractual and part-time workers, counted on any day during the year. Founders who count only permanent employees sometimes discover during an inspection that they crossed a threshold months ago.
And if you've read that PF now applies at 10 employees, that's incorrect. The Ministry of Labor and Employment explicitly denied those reports; it remains 20. That wrong figure still circulates in blog posts.
The timing that creates retrospective liability
Registration is due within 30 days of crossing a threshold. Miss it and you're in violation from day 31, and contributions are owed from the date you crossed, not the date you registered. Late registration means retrospective dues plus interest.
This is the strongest argument for buying slightly before you need to. Software that tracks headcount and warns you turns a liability into a calendar reminder.
The state layer
Professional Tax is a state tax, capped at ₹2,500 per person per year by the constitution. Several major states, including Delhi, Haryana, Uttar Pradesh, and Rajasthan, don't levy it at all. Others do, with different slabs.
The rule that matters: PT follows where the employee physically works, not where your company is registered. Hire a remote developer in Bengaluru, and you have Karnataka obligations, even with no office there. One employee triggers registration.
That's it. Five paragraphs, and you now know enough to ask questions a vendor has to answer properly.
The Questions That Separate Real From Marketing
Ask these. The answers are diagnostic.
"Show me an EPFO ECR file generated from my actual salary data." Every vendor says they handle compliance. This asks them to prove it. If they demo a sample file instead of generating one from your data, note that.
"How do you handle the basic-versus-allowance split for PF calculation?" PF computes on the basic salary plus DA. A vendor who can model your structure precisely will explain how. One who's vague is telling you something.
"An employee moves from Delhi to Bengaluru. What happens?" The correct answer involves a work-location state field that drives rule selection, with PT starting automatically. If the answer is "you update their address," the platform doesn't do multi-state.
"When Karnataka revises its minimum wage in April, who updates the software?" If the answer is you, you've bought a spreadsheet with a login.
"Can one person run this without a consultant?" Then ask how long setup takes. Days is right for your size. Weeks means it's built for someone bigger.
"What does it cost when I have 60 employees instead of 25?"
"Can I export all my data, and in what format?" Skipped constantly. Ask it before you sign, when you have leverage.
What to Ignore
Vendors will show you things designed to impress a founder who doesn't know what matters.
Performance management modules. Under 50 people, you know who's performing. You don't need a nine-box grid.
Engagement surveys and sentiment analysis. You can ask people how they're doing.
Recruitment pipelines, unless hiring is your bottleneck and you'd otherwise buy a separate tool.
AI-powered workforce analytics. With 30 employees, you have no meaningful sample. It's pattern-matching on noise.
Org chart visualization. You can see your org chart from where you're sitting.
"127+ features." A feature count is a marketing metric. You'll use eight.
None of these are worthless in principle. They're solving problems you don't have yet, and paying for them means navigating complexity every day for capability you won't use for three years.
What You Actually Need
Short list. All of it non-negotiable:
- Payroll with statutory calculations built in. PF, ESI, TDS, PT computed automatically, payslips generated, and the actual filing files produced. If you're doing any part in Excel afterwards, it hasn't solved the problem.
- Leave management matching your real policy. Not a simplified version. Test your actual policy, including carry-forward and probation rules.
- Attendance appropriate to how your team works. Field staff need mobile or GPS. Office staff might need biometric attendance. Fully remote teams mostly need leave tracking, not attendance.
- Employee self-service on mobile. This is where you get your time back. Employees checking their own payslips and leave balances stop asking you.
- Document storage with access control. Offer letters, PAN, bank details, and ID proofs. Not in your email.
- Data export. So this decision is reversible.
What It Should Cost
Indian HR software for SMEs runs roughly ₹20 to ₹200 per employee per month, with SME platforms commonly ₹48 to ₹150. Some price per organization, which usually favors you at small headcounts.
Rough calibration:
- Under 10 employees: free tiers exist covering payroll, leave, and compliance. Start there.
- 20 to 40 employees: per-organization pricing around ₹1,500 to ₹3,000 a month typically covers full payroll and compliance management.
- 30 to 50 employees: per-employee pricing lands around ₹4,000 to ₹8,000 a month on mid-market platforms.
Two traps. Payroll is frequently an add-on rather than included, so confirm the quoted price is the price. And a meaningful setup fee, at your scale, signals the product is built for larger companies.
How to Actually Test It
- Demos are designed to look good. Only your own data tells you anything.
- Run one full payroll cycle in the trial. Your real month, your real salary structures, your real employees. If it produces correct payslips, correct deductions, and correct filing files, you've found your answer. Everything else is a detail.
- Test the messy cases, not the clean ones. The mid-month joiner. The salary revision on the 15th. Eight days of unpaid leave. Someone on probation. A partial-month exit. Every system handles the standard employee. They differ entirely on these, and these are what your month contains.
- Have two employees use the self-service app. Not you. If they can find their payslip without asking, adoption will hold. If they can't, you've bought a support-ticket generator.
- Check the exit before the entrance. Export your data during the trial and look at what you get.
The Mistakes Founders Actually Make
- Buying for the company you'll be in three years. The cost isn't the money; it's navigating enterprise complexity daily for a capability you won't use until you choose a different platform anyway.
- Buying too late. The mirror image is more expensive. Crossing 20 employees unnoticed means retrospective PF with interest.
- Evaluating features instead of correctness. For a company under 50, correctness is the feature.
- Migrating mid-year without asking about it. Moving from spreadsheets mid-year means reconciling year-to-date figures. Get it wrong, and every subsequent payslip and Form 16 is distorted. Ask specifically how the vendor handles it. Better, start at a financial year boundary.
- Not checking whether one person can run it. If it needs an HR specialist and you don't have one, you've bought a hiring requirement.
- Skipping the trial because the demo looked good. The demo used their data.
The Decision, Compressed
Under 10 people: a free tier or basic payroll tool. Pick one that covers compliance properly so migrating later is a choice, not an emergency.
- 10 to 25: You're at or past the ESI threshold. Real compliance automation plus self-service. Per-organization pricing usually wins.
- 25 to 50: approaching or past PF territory, attendance is probably a genuine problem, and the admin load justifies a full HRMS. Investing properly here pays off, because migration later is painful.
- Any size: if it takes more than a few days to set up or needs a consultant, it's built for someone else.
Conclusion
You're not buying features. You're buying the correct statutory calculation, plus a system one non-specialist can run. Learn five facts, and you can evaluate properly: PF at 20 employees, ESI at 10, thresholds count contractors and part-timers, registration is due within 30 days with liability running from the crossing date, and professional tax follows where the employee physically works. Then make vendors prove it. Generate an ECR file from your data. Explain the basic-versus-allowance split. Show what happens when someone moves states. Run one full payroll cycle in the trial with your messy cases, not their clean ones. Skip the performance modules and the AI analytics. Buy for the company you have. And confirm you can get your data back out before you put it in.
