Blueprint
How to choose HR software: the Suite-or-Stack Test
Choose HR software by counting the jobs you will actually run in software this year, usually two or three for a small team, then price a suite against separate tools at your real headcount. On published 2026 prices, the same three jobs cost a ten-person team about $13 or $49 per employee per month depending on how well each pricing model fits. Give payroll its own decision, and check integrations and exit paths before signing anything annual.
The top-ranking guides for this search are written by companies that sell HR software. We checked in August 2026: every organic result on the first page came from a vendor, the number-one result was a white paper dated 2019, and all three of the guides we read end to end recommend their own vendor’s product and put no dated price next to any step of the advice.
This page takes the other route. The short version of how to choose HR software: count the jobs you will actually run in software this year, give payroll its own decision, price a suite against separate tools at your real headcount, and check the seams between tools before you sign. We call that sequence the Suite-or-Stack Test, and the rest of this guide walks it end to end, using the tools we have tested on real accounts and the dated prices we track by vendor. One scope note: prices are US list prices, captured between June and August 2026; the method travels anywhere, the dollar figures do not.
”HR software” is several bills wearing one name
HR software is not one product. It is at least five distinct jobs, each with its own market, its own pricing shapes, and its own failure modes. A common first mistake is comparing a suite that does four of them against a tool that does one, as if they were the same purchase.
| The job | What it covers | The pricing shapes on offer (published 2026 prices) |
|---|---|---|
| Core HR records | Employee files, time off, documents | Per-employee tiers from $0.84/mo (Zoho People, annual) to $25 (BambooHR Elite); flat floors like BambooHR’s $250/mo at 25 or fewer employees |
| Payroll | Pay runs, tax deposits and filings | A near-standard $35–49 base + $5–6 per person for full service; accounting bundles (QuickBooks, from $88 + $7); quote-only holdouts |
| Hiring (the ATS) | Job posts, candidate pipeline, offers | Free tiers, per-user seats ($15–19), flat plans ($157–189/mo), per-job-posting plans ($1,000/yr) |
| Onboarding | Offer paperwork, W-4/I-9 collection, first-week tasks | A tier feature of core HR or payroll in every tool we price, not its own bill; the tier it hides in varies by vendor |
| Global hiring | Employing people abroad | Employer-of-record fees from $599 per employee per month, a different budget conversation entirely |
Across the tools we price, two of these five do not arrive as their own bill: onboarding ships inside core HR or payroll tiers, and global hiring only applies if you employ abroad. That leaves at most three separate purchases for a small US team, and often fewer. Which ones you actually need is the whole question, and it is the step the vendor guides skip; the three we read all answer it the same way: all of them, ours.
The Suite-or-Stack Test
The test has four steps. It exists because the advice in the vendor guides, list your requirements, set a budget, demo a shortlist, quietly assumes you are buying one product. The real decision for a small team is earlier and bigger: one suite, or a small stack of single-job tools.
Step 1 — list the jobs you will run in software in the next twelve months. Jobs, not features. Payroll, hiring, records and time off, onboarding, performance. Next to each, write what handles it today and how many hours a month it costs you. A job that costs two hours a month in a spreadsheet has not earned a subscription; a job that eats a day a month, or carries legal deadlines, has. Twelve months is the honest horizon: a suite bought for the hiring push you might do next year is weight today. If you want a longer map of what deserves software at all, our guide to which HR processes are worth automating sorts that question job by job.
Step 2 — give payroll its own decision. Payroll is the one HR job with statutory deadlines and penalties behind it: the paycheck that must land, and the tax deposits and filings the tool carries out while the legal liability stays with you, the employer. It deserves its own elimination-first pass, which we wrote up separately as the Hardest-Constraint Test for payroll: find the constraint that eliminates tools (contractors, multiple states, tips, garnishments), then compare survivors on cost at your headcount, the year-two price, and reachable support. Decide payroll first, because the payroll choice constrains everything else: the winner may bundle enough light HR (self-service records, time off, onboarding) to strike jobs off your Step 1 list. Gusto, which we run our own pay cycles on and scored 8/10, includes employee self-onboarding from its base $49-plus-$6 plan; that fact alone removes a line from many small-team stacks.
Step 3 — price both shapes at your real headcount. Take your Step 1 list and price it twice: once as a suite, once as separate single-job tools, both as a full year at your actual headcount. Not the sticker: the year, with add-ons and billing terms in. The arithmetic regularly surprises people, and the next section shows why.
Step 4 — check the seams before you sign. A stack lives or dies on what flows between the tools; a suite lives or dies on what it lets you take out. Both checks belong in the demo, not in the first support ticket. The seams checklist below covers what to verify.
Suite or separate tools: what the arithmetic says
Neither shape wins in general; the arithmetic decides case by case, and it swings hard. One honesty note first: a true ten-person suite price is often quote-gated (Rippling and Paychex publish no list price at all), so what we can price openly is the force that decides the suite-or-stack outcome either way: pricing-model fit. We priced a ten-person company two ways on published 2026 prices, same three jobs: the shape-fit stack (a $0.84-per-user records tier, base-plus-per-person payroll, one $15 recruiter seat) came to about $13.24 per employee per month, and the same jobs on models sized for other shapes, led by a flat suite-style floor priced for twenty-five heads, came to $49.13. Nearly four to one, before a single feature is compared.
The mechanism is floors and units. A flat $250-a-month minimum, which is what BambooHR’s floor is at 25 or fewer employees, works out to $25 a head at ten people and $10 a head at twenty-five. The per-employee sticker only means something at the headcount where the floor dissolves. Divide every flat price by your actual headcount before comparing anything.
The flip side deserves equal print: suites win real cases. At twenty-five people, that same $250 floor meets BambooHR’s own published $10 per-employee rate, and the flat plan stops charging you for being small. A team that genuinely runs four or five of the jobs, not two, saves real integration effort under one roof, and BambooHR, to keep the example, includes onboarding packets on every plan rather than hiding them in a top tier. And a company already living inside one vendor’s wider ecosystem can pay a suite price that a standalone buyer should refuse. The test is not “suites are a trap”; it is that the suite must beat the stack on your Step 1 list, at your headcount, priced for a year; in the ten-person example above, the mismatched shapes lost that comparison by nearly four to one.
One more asymmetry worth pricing: renewals. A stack spreads renewal risk across small, replaceable subscriptions; a suite concentrates it in one negotiation where the vendor holds your data. Renewal uplifts of 8 to 15% are reported for Greenhouse contracts, and we documented Gusto’s base fee moving from $40 to $49 between two dated captures in 2026. Whatever shape you pick, get the year-two price in writing.
The seams: what must flow between your ATS, onboarding and payroll
An “integrates with” logo on a pricing page tells you a connector exists, not what it moves. For a small stack, these are the four seams we check first, each with a concrete thing to verify in the demo, with your own test data, before you sign.
| Seam | What must flow | What to verify in the demo |
|---|---|---|
| ATS → onboarding | The hired candidate’s name, role, salary and start date become a new-hire record | Convert a test candidate to hired. Does the data carry over, or does someone re-type it? |
| Onboarding → payroll | W-4, I-9 and bank details land in the payroll record | Complete a test packet. Where the forms live matters: Homebase, for example, only carries its W-4/W-9/I-9 packets on the $120-per-location All-in-One plan (captured Aug 2026), not the $30 tier its pricing page leads with |
| Time → payroll | Approved hours reach the pay run without a CSV export | Approve one timesheet and open the draft pay run |
| Payroll → books | Each pay run posts to the ledger as journal entries | Run a test cycle and look at the journal entry, not the marketing page |
The payroll-to-books seam is currently being bundled away by the vendors themselves: QuickBooks now sells payroll only inside its accounting bundles (captured 11 Aug 2026), and Xero embeds a Gusto-powered payroll at $36 plus $6 inside its accounting plans, cheaper than Gusto’s own $49-plus-$6 direct price if you already want Xero’s books (captured 20 Aug 2026). We price that whole one-bill-or-two decision separately.
And there is a fifth, often-missed seam: the exit. Ask how each tool exports your data the day you leave, and get specifics. On the hiring side we found a sharp example: Breezy’s free plan carries a 30-day candidate-data window alongside its one active position (per its help centre, checked Aug 2026), so the export question is not theoretical even at $0. For payroll, the question is year-to-date figures mid-year; for records, it is documents and history, not just a names list.
Questions to ask vendors before choosing HR software
Every question below is backed by a wall we documented, on a dated capture or a tested account. Ask them exactly, in writing where it matters.
- “What will this cost at renewal, in writing?” Prices move: we documented one base fee rising $40 to $49 within months, and 8 to 15% renewal uplifts are reported at the enterprise end. A refusal to name the year-two price is an answer.
- “Which of the features you just showed me are add-ons, and what is the add-on price list?” JazzHR’s $1,000-a-year plan meters extra job posts at $9 each per month and sells texting, offers and reporting separately; Workable’s AI credits run out and cost $600 per 5,000-credit top-up, and they expire after a year; Zoho Recruit sells its client portal and video interviews as line items. The add-on list is usually harder to find than the headline, which is itself information.
- “What does growth cost: one more hire, one more state?” Per-person fees compound with every planned hire, and Patriot, to name one, files one state free and charges $12 a month per extra filing state. Price your December headcount, not your January one.
- “Is your pricing published anywhere?” Rippling, Paychex, ADP Run and Greenhouse published no list price as of our mid-2026 captures, and Tellent gates its pricing behind a demo by employee count. Quote-only is not automatically expensive, but it takes away your ability to comparison-shop in an afternoon, and that ability is most of the bargaining power a small team has.
- “How do we export everything if we leave?” Candidates, documents, year-to-date payroll, history. Ask for the export path on your plan, then test it during the trial. The demo is the negotiation; the divorce is not.
- “Who answers support on the plan I am buying?” Not the top plan: yours. Chat bot, human chat, email, phone. Then measure it once during the trial with a real question. The answer you get while they are courting you is the best answer you will ever get.
Run the trial like a test, not a tour
A vendor demo follows the vendor’s script. A trial should follow yours: pick the one real workflow the tool is being hired for and run it with real stakes, the way we test tools before scoring them. Post one real job and process the real applicants that arrive. Enter one real employee record and request real time off against it. If it is payroll, run a parallel cycle against your current numbers before you switch anything. One real workflow surfaces more than two weeks of feature tours, because your edge cases show up and the scripted ones do not.
While the trial runs, collect the two numbers no pricing page prints: hours to a human on your plan’s support channel, and minutes to a full export of your own test data. Both are cheap to measure in week one and expensive to discover in month eleven.
Common mistakes when choosing HR software
Buying the suite by default. Modules you will not run are not free, they are weight: setup surface, admin surface, renewal surface. Buy your Step 1 list, not the category.
Comparing stickers across pricing models. A flat floor, a per-employee rate and a per-job plan cannot be compared as printed. Multiply everything out at your real headcount, for a year, at the billing term you will actually pick; on plans we track, the annual-versus-monthly delta alone runs from about 17% (Breezy’s $189 plan billed at $157) to about 21% (Manatal’s $19 seat billed at $15).
Sizing the ATS seat like a permanent job. Hiring is spiky for small teams. A team that hires one role at a time can sit on a free tier and pay nothing between hires, or on a per-posting plan sized for low volume; steady hiring earns a paid seat. Our small-business ATS round-up prices those shapes against each other, tested picks first. (And if you hire for clients rather than for yourself, you are shopping a different category: start from the agency-side ATS picks.)
Skipping the exit check. Every question in the seams table is easy in the demo and hard in month eleven. If you only verify one thing from this page, verify the export.
Trusting undated prices. HR software prices moved this year: we captured one vendor’s base fee rising $40 to $49, and another’s own pricing FAQ flags a recent base-price change by state. A price without a date next to it, on any site including review sites, deserves suspicion; it is why every number we publish carries one.
The one-page version
Choosing the right HR software is a single sequence. Count the jobs you will actually run in software this year; most small teams find two or three. Decide payroll first and separately, because it carries the legal deadlines and its winner may cover other jobs in passing. Price a suite against a stack at your real headcount, for a full year, floors divided by your actual heads. Then check the seams: what flows between tools, and what comes out the day you leave. The feature grid can wait until the shortlist is down to survivors; the shape of the bill cannot.
FAQ
What is the difference between HRIS, HCM and HRMS?
In vendor marketing, HRIS usually means the core employee-records system, HCM adds the talent lifecycle around it (recruiting, performance, succession), and HRMS often means an HRIS with payroll attached. In practice the three labels are applied inconsistently, sometimes to the same product. Ignore the acronym and read the module list: you are buying a specific set of jobs, not a category name.
Which HR software is best for a small business?
There is no single best, and any page that names one without asking what you need is guessing. The honest version: match each job to the pricing shape that fits it. On published 2026 prices, a ten-person team covers employee records, payroll and one recruiting seat for about $13 per employee per month bought shape-first, and about $49 buying the same jobs on mismatched models. The tools we have tested hands-on carry dated scores in our reviews; the rest we price but do not rank.
How much does HR software cost in 2026?
On published 2026 prices: core HR runs from $0.84 to about $25 per employee per month, full-service US payroll adds a $35 to $49 monthly base plus $5 to $6 per person, and an ATS runs from free to a few hundred dollars a month depending on its model. Our worked ten-person stack lands near $13 per employee per month; the same jobs on mismatched models reach $49. Every number carries a dated source in our cost guide and pricing index.
Do I need HR software at all under ten employees?
Payroll is the one job that usually justifies paid software early, because tax deposits and filings carry statutory deadlines and penalties; on published 2026 prices, full-service payroll runs roughly $60 to $80 a month at five people. Records, time off and hiring can live in spreadsheets and inboxes until the volume hurts. When a job starts costing you hours every week, or risking mistakes with legal weight, that job has earned software. Buy job by job, not all at once.
