ATS Implementation: A Phase-by-Phase Rollout Plan

The numbers behind this

How long does an ATS implementation take?

Company size and data mess decide the calendar more than the vendor does. A single recruiter switching tools can be live in a long afternoon: create the account, post two roles, import a spreadsheet of active candidates, done. A twenty-person talent team with an HRIS connection, custom scorecards, and four years of history usually needs six to twelve weeks. Anything with multiple legal entities, works councils, or a security review will run a quarter or more, and most of that time is waiting on other departments rather than software work. Two variables inflate every estimate: how clean your existing records are, and how many people must approve a decision. Budget slack for both. Set the go-live date after discovery, never before, and pick a quiet week in your hiring calendar. Teams still shortlisting can start from a guide to choosing an ATS before scoping the project.

What happens in each phase of an ATS rollout?

Nine phases, run in order, with two that people try to skip and regret. Discovery maps your current process honestly, including the workarounds nobody documented. Configuration builds pipelines, stages, scorecards, email templates, and permissions. Migration moves the records worth keeping. Integrations connect calendar, email, HRIS, and job boards. Workflow and template build is where automation rules get written and tested against real scenarios. Training splits by audience, because recruiters and hiring managers need different sessions. The pilot runs two or three live requisitions end to end in the new system while the old one stays available. Cutover freezes the legacy tool and redirects careers-page traffic. Optimization starts two weeks later, once you know what actually annoys people. The skipped ones are always pilot and optimization, and skipping either is why teams end up back in spreadsheets. The table below sets out who is involved and for how long.

What candidate data should you actually migrate?

Migrate less than you think. Active candidates in live pipelines, open requisitions, hired-employee records, and anything with a legal retention obligation must come across. Everything else deserves a decision. Rejected applicants from four years ago, duplicate profiles, resumes attached to roles that no longer exist, and notes written in shorthand nobody can read are dead weight that will pollute search results and skew your first reports. Export the old database, deduplicate on email, and agree a cut-off date. Two years of history is plenty for most teams. Keep a read-only archive of the rest rather than importing it. Check consent and retention rules before moving personal data between systems, especially across borders. Run the migration twice: a test load you inspect field by field, then the real one during a freeze window. Reconcile counts afterwards. Switch support is outlined on the migrate to Pitch N Hire page.

• ▸ Bring across: active candidates, open requisitions, hires, and anything legally retained
• ▸ Leave behind: stale rejections, duplicates, orphaned resumes, and unreadable shorthand notes
• ▸ Deduplicate on email address before import, not after
• ▸ Agree a cut-off date; two years of history covers most teams
• ▸ Test load first, inspect field by field, then run the real load in a freeze window
• ▸ Reconcile record counts against the export before anyone declares migration finished

Who needs to be on the internal implementation team?

You need four roles filled by name, not by department. A project owner from talent acquisition makes decisions and holds the timeline; without one, the vendor ends up designing your process. A system administrator learns the configuration properly and becomes the person everyone asks, which is a real time commitment rather than a title. A data owner handles the export, the deduplication, and the reconciliation, and should be someone who enjoys spreadsheets. An executive sponsor removes blockers, especially with IT and legal, and signs off the go-live. Add one respected recruiter and one hiring manager as pilot users, because early adopters who are trusted internally convert doubters faster than any training deck. Estimate ten to fifteen hours a week for the owner and administrator during the build. Underestimating this is the single most common planning error, and it shows up as a slipped date nobody predicted.

Why do ATS rollouts fail, and how do you avoid it?

Rollouts rarely fail on technology. They fail because the new system encodes a process the team never agreed to, so people quietly keep working the old way and the data goes stale within a month. The second killer is rebuilding your broken process faithfully in new software, complete with the fourteen-stage pipeline nobody follows. Implementation is the one moment you get to simplify. The third is a big-bang cutover with no pilot, where every problem surfaces at once in front of live candidates. The fourth is losing candidate communication during the switch; 36% of candidates never hear back after applying, according to Talent Board's CandE research, and a messy migration makes that worse. Fifth is having no named owner, so decisions queue. Fix them by agreeing the process before configuration, cutting stages, piloting on real reqs, and assigning one name to every open question.

How should you train recruiters and hiring managers?

Train by job, and keep it short. Recruiters need ninety minutes on pipelines, search, bulk actions, templates, and reporting, followed by supervised use on a live role. Hiring managers need fifteen minutes and one job aid, because anything longer will not be watched: review candidates, leave a scorecard, approve or reject, done. Coordinators need scheduling and calendar edge cases. Executives need the dashboard and nothing else. Record every session and keep a searchable one-page reference in the tool your company already uses. Then answer the question people are actually asking, which is what changes for me on Monday. Nominate two champions per team who get early access and become the first line of support. Run a drop-in clinic each week for the first month. Adoption is a habit problem, not a knowledge problem, and structured interview scorecards and questions give managers a reason to log in.

What should happen in the first 30 days after go-live?

Week one is triage. Watch the queue daily, fix broken automation immediately, and keep a visible list of reported issues with owners and dates so people believe reporting them is worth it. Confirm every job posting syndicated correctly and that applications are landing, because a silent form failure costs candidates you never learn about. Week two, check the candidate side: apply to your own job on a phone and count the steps. Long forms lose people, and 51% of candidates abandon applications that take too long, per SHRM. Week three, review adoption by user, not by team average, and call the people who have not logged in. Week four, run your first funnel report and compare it against your baseline time to fill, which averages around 45 days according to SHRM's 2025 Benchmarking Report. Then cut one stage. Tracking time to fill improvements early sets the habit.

ATS implementation phases, owners, and typical duration by company size