HR Tech Stack: The 5 Layers, What to Buy First, and How to Build by Company Size

An HR tech stack is the set of software tools a company uses to manage its people: employee records, onboarding, payroll, compliance, and (as the company grows) hiring, performance, and analytics. At an enterprise, this stack can include a specialized tool for nearly every function. At a small business, it should include 2, maybe 3.

The problem with most HR tech stack guides is that they describe the enterprise version: HRIS (human resource information system) plus ATS (applicant tracking system) plus payroll plus performance management plus engagement surveys plus learning management plus workforce analytics. That is the right stack for an enterprise with a dedicated HR team.

The same stack is the wrong one for a small business where the founder handles HR between product meetings. I build HR software for small teams, and my advice to any founder is to buy the smallest stack that covers the work.

This guide covers the 5 layers of an HR technology stack, what to buy first, how the stack should evolve as you grow, and the mistakes that lead companies to either under-invest (spreadsheets until something breaks) or over-invest (enterprise tools at startup scale).

What Is an HR Tech Stack?

An HR tech stack (also called an HR technology stack or HR software stack) is the combination of software tools that a company uses to manage its human resources functions. It spans the employee lifecycle: from hiring (applicant tracking) through onboarding (workflows and document collection) to employment (records, payroll, performance) to departure (offboarding, exit documentation).

The stack metaphor comes from software engineering, where a "tech stack" refers to the layers of technology that work together to run an application.

In HR, each layer handles a different function, and the layers connect through integrations (data flowing from the HRIS to payroll, from the ATS to onboarding, from performance reviews to analytics). The quality of these integrations matters as much as the quality of the individual tools.

The 5 Layers of an HR Tech Stack

Every HR tech stack, regardless of company size, is built from these five layers. The difference between a small business and an enterprise is not which layers exist. It is how many layers are active and how sophisticated each one is.

The layers are listed in priority order. Layer 1 (foundation) should be implemented first because every other layer depends on it. You cannot run payroll without employee records. You cannot track performance without knowing who reports to whom. You cannot analyze workforce data without a database to analyze. Add each remaining layer only when going without it costs you more than adopting it would.

Layer 1: The Foundation (and Why It Comes First)

The foundation layer is the non-negotiable. It is the system that stores every employee record, automates onboarding for every new hire, and manages compliance documents with e-signature.

The same platform draws a visual org chart from those records and gives employees self-service access to their own information. Without a foundation layer, employee data scatters across spreadsheets, filing cabinets, email, and the founder's memory.

A platform like FirstHR covers the entire foundation layer: AI-powered onboarding wizard, e-signature, HRIS with employee profiles, visual org chart builder, training modules, and self-service portal. The cost is $98 per month flat for 10 active team seats, or $198 per month for 50. This is the foundation that everything else builds on.

According to Gallup (2018), only 12% of employees strongly agree that their organization does a great job onboarding new employees. The foundation layer is where onboarding lives, so it is the layer to get right first.

What the Five Layers Leave Out

The five-layer model describes the HR functions every company eventually touches. It does not describe every piece of software you may be legally or operationally required to run. Four functions sit outside the core layers, and whether you need them depends on how you employ people rather than on how many people you employ.

Time and attendance is the one that most often gets skipped and most often costs money later. Department of Labor Fact Sheet 21 puts the duty on the employer rather than on a timekeeping vendor: hours worked each day and each week for every nonexempt employee, in whatever form you choose to keep them.

Wage and hour claims are decided on those records. When the employer has no reliable record of hours worked, the employee's reasonable estimate tends to carry the day under the Supreme Court's ruling in Anderson v. Mt. Clemens Pottery Co. (1946). Timekeeping that amounts to a group text and a manager's memory is therefore a liability sitting outside your stack, not a cost you have avoided.

Rounding rules deserve the same attention. Federal law (29 CFR 785.48) permits rounding only when it is neutral over time, meaning it cannot consistently favor the employer. Some states have narrowed or effectively eliminated it, so a rounding setting your vendor turned on by default is worth checking against your own state's rules.

Sample Stacks by Company Size

The table below shows which layers are active at each stage of growth and how the monthly cost is built.

Payroll is priced as a monthly base fee plus a per-employee fee. Published list prices for full-service small-business plans, read in October 2026, start at $29 to $50 a month plus $5 to $7 per employee on entry plans and reach $180 plus $22 per employee on premium tiers. Some large providers publish no prices and quote only through sales.

Notice the pattern: the foundation layer and payroll are constant across every size. Payroll is the layer small businesses most often hand to a specialist provider rather than build themselves. What changes is how many additional layers are active and how sophisticated each tool is within a layer.

A small business does not need a $500 per month ATS. A spreadsheet or free job board handles fewer than 10 hires per year. The ATS becomes worth its cost when hiring volume exceeds 10 positions per year, and the cost of manual tracking (lost candidates, inconsistent screening, slow response times) exceeds the tool cost.

How to Build Your HR Tech Stack

Build the stack in five steps, in this order: foundation, payroll, free tools for early hiring, performance tools, and analytics. The table pairs each step with the mistake to avoid at that stage.

The principle: add layers when the pain of not having them is greater than the cost of adding them. If you have never missed a compliance deadline, your document management is working. If candidates are not falling through the cracks, your hiring process is working. Do not fix what is not broken. Invest where it hurts.

Research reported by SHRM (2017) shows that up to 20% of employee turnover occurs within the first 45 days of employment. If early turnover is your pain point, invest in the foundation layer (onboarding), not analytics.

What a Digital HR Strategy Means at This Size

A digital HR strategy is the decision about which HR work stops being manual, in what order, and how you will know it worked. At enterprise scale that becomes a multi-year transformation program with a budget and a steering committee. In a company with no HR department, it is the sequence above plus two things that sequence does not cover.

The first is adoption. A platform that managers never open leaves the founder doing the work by hand anyway, so every tool needs one named owner, a date the old process stops, and a plan for retiring the spreadsheet it replaced. Running the manual process alongside the new one indefinitely is a common way for a rollout to quietly fail.

The second is measurement. Pick two numbers before you buy: the hours per week the process consumes today, and one outcome the tool is supposed to move, such as paperwork completed before day one or turnover inside the first 90 days. Look at both a quarter after launch. If neither moved, the tool was not the fix.

How Data Actually Moves Between Tools

"Integrates with payroll" is a sentence that hides five very different things. Before you accept it from a vendor, ask what kind of connection it is, which direction data travels, and how often it runs. Those three answers determine whether the integration saves you work or just relocates it.

The rule that keeps a multi-tool stack coherent is one field, one owner. Decide, field by field, which system is authoritative, and make every other system a copy.

In practice, the HRIS owns identity data (legal name, hire date, department, manager, employment status), payroll owns compensation history, tax withholding, and bank details, the benefits carrier or broker owns enrollment elections, and the ATS owns nothing after the offer is accepted. When two systems both claim to own pay rate, you will eventually pay someone the wrong amount and spend a day working out which record was right.

Direction matters as much as ownership. A termination entered in the HRIS should stop the next payroll run and trigger a benefits termination that starts the COBRA notice clock (the countdown to offering continued health coverage). A raise entered directly in payroll, bypassing the HRIS, leaves your records showing the old number the next time you benchmark compensation or defend a pay equity question.

The fix is to pick the entry point for each event type, write it down, and hold to it. Most stack failures are not integration failures. They are two people entering the same change in two different places.

Write it down here, before the next tool goes in. The first tab settles ownership field by field and how each copy is kept current; the second settles where each event gets entered and what it has to set off downstream. The seeded rows are the split described above, and the blank ones are the fields your own stack argues about.

HCM Integration in a Small Stack

HCM integration, meaning how a human capital management (HCM) suite connects to the rest of your tools, is the same question one tier up. That kind of suite bundles the HRIS, payroll, benefits, and talent layers under one vendor, so the connections that matter point outward: to accounting for labor cost, to benefits carriers for enrollment, and to identity management for provisioning and shutoff on the last day.

A suite is sold as the end of the integration problem, and it relocates that problem rather than removing it. Ask a suite what you would ask a connector: which fields the modules genuinely share, which ones still move by file transfer, and whether accounting can pull labor cost without someone exporting it by hand every month.

The difference between an HRIS and an HCM is worth settling before a demo settles it for you, because for most small businesses a full suite is more platform than the work actually requires.

Migrating Off Spreadsheets Without Losing Records

The move from spreadsheets to a real foundation layer is usually described as a one-day import. It is closer to a two-week project at small-business scale, and most of the work is cleanup rather than data entry.

The spreadsheet has three versions, two of them on someone's desktop; job titles do not match what is in the offer letters; four people have no recorded start date. Migration is the moment all of that surfaces.

A sequence that works: first, freeze the spreadsheet and declare a single version authoritative so changes stop landing in the copy nobody is migrating. Second, reconcile the employee list against payroll, because payroll is the one system that is definitionally complete. Anyone being paid is an employee, and anyone in the spreadsheet who is not in payroll needs an explanation.

Scope and proof come last. Third, migrate active employees only, and archive terminated employees as records rather than trying to reconstruct their full profiles. Fourth, run one full pay cycle with both systems in view before you retire the old process.

The reconciliation is the step worth doing on paper rather than in your head, because every mismatch it surfaces needs an explanation and a name against it before anything gets imported.

Work the reconciliation one row per person, from the payroll list down, and record two things against each row: why the person appears in one list and not the other, and where the old file went. The employee directory guide has a reconciliation worksheet built for exactly this pass, and it works the same whether the destination is a directory or the whole foundation layer.

Retention is the part people get wrong. Migrating systems does not reset any clock and does not let you discard the old files. Under 8 CFR 274a.2, an I-9 is kept for three years after the date of hire or one year after employment ends, whichever is later.

The I-9 is also best kept separately from the personnel file, so an audit does not expose everything else. Under the FLSA, payroll records are kept three years, and the records that wage computations are based on, including time cards, are kept two.

Several other rules run longer or shorter, and some states impose their own periods, so the practical answer is to keep the old records for the longest applicable period rather than deleting anything at migration.

If you store I-9s electronically in the new system, the electronic storage has requirements of its own: controls that maintain the integrity and accuracy of the record, indexing that lets you retrieve a specific form, an audit trail showing when a record was created or changed, and the ability to produce a legible copy on request.

Changing software is not a reason to re-collect I-9s from existing employees. Completing a new form when none is required creates its own problems, including the appearance of re-verifying employees you had no basis to re-verify.

When a PEO Replaces Most of the Stack

There is an alternative to assembling layers, and small employers should price it before they build: a professional employer organization (PEO). A PEO enters a co-employment relationship with you. It becomes the employer of record for payroll tax purposes, runs payroll, files employment taxes, sponsors or administers benefits, and usually carries workers' compensation under a master policy.

You keep day-to-day direction of the work, hiring and firing decisions, and the business itself. In stack terms, a PEO collapses layers 1, 2, and most of benefits administration into one relationship, and gives you its platform whether or not you like its platform.

Tax liability is the piece to confirm rather than assume. The IRS holds employers ultimately responsible for withheld income tax and for both shares of Social Security and Medicare, and relieves that liability, in certain situations, only for customers of a certified PEO (CPEO).

Run the comparison in one unit before deciding. A percentage-of-payroll quote and a per-employee quote are not comparable until you convert both to annual dollars at your actual wage levels.

As an illustration, a quote of five percent of gross payroll on a $2 million annual payroll is $100,000 a year. The same headcount on a software stack plus a broker is a different order of magnitude, but that number does not include the benefits pricing advantage, the workers' compensation policy, or the state registrations. Compare what is inside each number, not the headline.

One thing a PEO does not do is take away your status as an applicable large employer. If you averaged 50 or more full-time employees including full-time equivalents during the prior calendar year, the IRS treats you as the applicable large employer under the ACA employer mandate, not the PEO.

The responsibility for offering compliant coverage therefore rests with you even when the PEO prepares and files the forms. The same logic applies to most obligations: a PEO changes who performs the work, not who ultimately owns it. The certified-PEO tax rule above is an exception.

Common HR Tech Stack Mistakes

Six mistakes are worth avoiding. They run from buying too much too early to waiting until a spreadsheet breaks, and each row pairs the mistake with what to do instead.

The most expensive mistake is not choosing the wrong tool. It is using too many tools with no integration. Every disconnected tool creates a data silo where employee information lives in isolation. When data is scattered across 5 tools, no single tool has a complete picture, and the founder spends hours reconciling information that should flow automatically.

Consolidation is the answer, and the stakes are higher than administrative tidiness. According to Gallup (2024), 42% of employees who voluntarily left their organization say their manager or organization could have done something to prevent it. Some of what they needed sits in the records and reminders your stack is supposed to hold.

Frequently Asked Questions