A 30-60-90 day onboarding plan works best when the numbers are review points, not automatic deadlines for declaring someone "adapted." What should drive the transition from one phase to another is the readiness to assume a greater level of responsibility, demonstrated by combined evidence before work begins. Thus, the 30, 60 and 90 days organize the conversation, while the role criteria organize the decision.
This design avoids two common problems. The first is to transform onboarding into a sequence of administrative tasks that ends without showing whether the person actually learned to work in that context. The second is to use check-ins as a surprise assessment, demanding expectations on the 60th or 90th day that were never explained. A good plan makes visible, from the beginning, what needs to be learned, which deliveries will be made with supervision, what autonomy is expected and who should support each step.
The 30-60-90 plan should control progression, not calendar
The most useful logic is to start with the desired state of the function and work backwards. Before filling out any model, the manager needs to answer what a minimally productive and confident person can do without constant help, which decisions still require validation, which mistakes are acceptable during learning and which risks require direct supervision.
With this defined arrival point, 90 days are no longer a promise of full productivity. In simple and repetitive functions, operational autonomy may arrive sooner. In complex, regulated, technical or leadership roles, it may take longer. The plan remains useful because it organizes the progressive reduction of dependency, but does not force an artificial date for something that depends on the complexity of the role, the person's repertoire and the quality of support received.
The 30-60-90 day structure appears frequently in onboarding guides because it helps separate three different movements: understanding the context, contributing safely, and assuming increasing responsibility. The real gain is in this progression. If the document just swaps one task list for three smaller lists, the company still doesn't know what it means to be ready to move forward.
The 30, 60 and 90 day matrix: learn, deliver with supervision and act autonomously
Days 1 to 30: learn the system before charging speed
In the first 30 days, the priority should be to reduce uncertainty. The person needs to understand the role of the area, the priorities, the processes that support the work, the tools, the risks, the stakeholders and the way decisions are made. You also need to know what you shouldn't do alone yet.
The goal of this phase is not to "get to know the company" in a generic way. It is about building enough knowledge to interpret real job situations. A better criterion than completing training is being able to explain the workflow, find reliable information, recognize when to ask for help, and perform a low-risk task with close guidance.
The manager must reserve time for frequent conversations, because initial learning does not just happen in courses and documents. Surveys and onboarding guides from organizations such as Microsoft and Gallup reinforce the importance of the manager's active participation and connections with colleagues during adaptation. In practice, this means that access to context and feedback needs to be part of the plan, and not rely on the person "jumping around" to figure out who knows what.
Days 31 to 60: transform knowledge into supervised delivery
Between days 31 and 60, the question changes. Instead of “what does this person already know?”, it becomes “what can they produce by applying this knowledge with a safety net?”. It is the appropriate phase to assume bounded deliveries, make lower-risk decisions, deal with known exceptions and receive feedback on quality, prioritization, communication and method.
Supervision still exists, but its format must change. At the beginning, the manager can explain each step. Then, it goes on to review decisions and results. The criterion for advancement is not the absence of doubts, but rather the ability to use the process, ask for help at the right time, incorporate feedback and repeat a delivery with a predictable level of quality.
Good evidence of readiness at this stage is observable. It could be completing a real cycle without critical rework, conducting a conversation with a client with subsequent review, producing a report with a manager's conference or delivering a small functionality respecting team standards. The concrete example depends on the position; The principle is always the same: real work, controlled scope and criteria known before execution.
Days 61 to 90: extend autonomy without withdrawing support too early
On days 61 to 90, the focus is on checking whether the person can take on a stable portion of the work with less dependence. Autonomy does not mean working in isolation or failing to ask for help. It means knowing how to conduct routine work, recognize limits, escalate relevant exceptions, prioritize within scope and respond to agreed deliverables.
The most common mistake at this stage is confusing autonomy with volume. A person can produce a lot and still rely on frequent corrections, or produce less because they have taken on more complex tasks and demonstrate excellent judgment. Therefore, the readiness criteria needs to consider quality, consistency, decision-making, communication and the ability to recognize risks, and not just quantity delivered.
Day 90 should close an integration cycle, not automatically declare that development is over. The expected result is clarity about the level of autonomy achieved, the points that still require support and the goals for the next period. When the role requires a greater curve, the plan can continue with new milestones without turning extension into a sign of failure.
How to write readiness criteria that really help you decide
Readiness criterion is an observable condition that shows that the person can assume more responsibility with acceptable risk. To work, it needs to describe behavior or outcome, indicate what evidence will be considered, and make clear the level of support allowed. "Being adapted" is vague. "Conduct process X from start to finish, using the current procedure and escalating defined exceptions, with review only at the end" is much more useful.
It is also important to separate criteria from activity. Attending three meetings, attending five trainings, or reading a manual may be necessary, but this does not prove mastery. Activity is a means. Evidence is what demonstrates that learning has been transformed into execution capacity. Therefore, the plan must record both what the person will do to learn and what they must demonstrate later.
The level of quality needs to be proportional to the moment. In the first month, understanding and accompanied execution are expected. In the second, consistency in known situations. In the third, autonomy within a defined scope. Demanding the same standard from an experienced person and a newcomer in the first cycle eliminates the pedagogical function of onboarding and tends to turn each check-in into a hidden test.
Who is responsible for what in onboarding
HR is responsible for architecture and consistency
HR must ensure that there is a common process, that access, documents, policies, mandatory training and integration resources are available and that managers know what their role is. It is also up to HR to create the model that helps record goals, responsible parties, evidence and check-ins without consolidating all functions into a single script.
HR should not be the sole owner of the new employee's success. He does not know all the decisions, priorities and quality standards of each position. When onboarding is
centralized only in HR, it is common for the process to work well until the institutional presentation and lose momentum precisely when the real work begins.
The manager is responsible for the criteria of the role and the reduction of dependency
The direct manager is the one who translates the function into results, limits and criteria. It defines which deliverables are appropriate for each phase, agrees what will be reviewed, ensures availability for feedback and decides when the person can expand the scope. Your job is not just to stick to the plan, but to remove obstacles and calibrate the difficulty of learning experiences.
This responsibility also requires making expectations visible. A goal shouldn't just appear at check-in. If the manager realizes that a skill will be necessary, he needs to communicate it beforehand, offer context and create an opportunity for the person to practice it. The plan should function as a learning and contribution contract, not as an archive of evidence used retroactively.
The new employee responds by actively participating in the process
The person hired also has responsibility. It should signal doubts, record impediments, ask for examples when a criterion is abstract, seek feedback on deliveries and gradually adopt the plan itself. This does not transfer the obligation to find out what the company expects on her own; it means participating in building clarity and demonstrating how you are using the support offered.
When there is a buddy, mentor or reference colleague, their role is to complement the manager. This person helps navigate informal practices, locate information, and understand the team's social context, but should not be used to replace leadership conversations about priorities, performance, or expectations.
30, 60 or 90 day check-in is not the same thing as a formal evaluation
The check-in serves to compare progress against the plan, review evidence, identify obstacles and adjust the next level of responsibility. A formal evaluation, when it exists in the company, has another function: to consolidate performance according to a process, period and specific performance management criteria.
As a conceptual reference, the United States Office of Personnel Management separates continuous performance monitoring, based on frequent measurement and feedback, from the moment of formal rating. This distinction is useful outside the North American public context because it shows why monitoring and evaluation need not be confused. Monitoring exists to correct the course early; formal evaluation summarizes performance within defined rules.
To avoid surprises, every 30, 60 or 90 day conversation should start with criteria that the person already knew. If a new expectation arises, it can enter the next cycle, but it should not be used retroactively to declare that the previous phase was unsuccessful. The manager must also record concrete examples, including when the result is positive, so that the feedback does not become a collection of vague impressions.
When a criterion has not been met, the first question must be causal. Was there a lack of knowledge, practice, access, clear priority, real opportunity to execute or minimum quality in delivery? The answer defines the next step. Sometimes it is necessary to reinforce training or reduce scope; in others, increase supervised practice or make the expectation more specific. Only after separating onboarding design flaws from performance difficulties is it possible to make a fair decision.
Time to productivity: measuring readiness for the role, not a feeling of adaptation
SHRM treats time-to-productivity as the time it takes a new hire to achieve defined performance indicators for the position. This formulation solves an important problem: productivity should not be a universal date nor a manager's perception. You need a reference point compatible with the job.
To use this metric, the company must first define which minimum set of results characterizes the expected productivity for that stage of the function. In customer service, it may involve handling routine cases with quality and knowing how to escalate exceptions. In an analytical role, it may mean closing a reporting cycle accurately and explaining relevant variances. In sales, it could be driving steps of an opportunity within the process and keeping the CRM reliable. The number of days only gains meaning after this state has been defined.
It is also useful not to confuse time to productivity with maximum performance. A new employee can reach a basic productive level and continue increasing speed, repertoire and impact for months. The indicator should mark when the person starts to contribute sustainably at the level expected for their internship, not when they reach the performance ceiling of someone who already dominates the organization.
Examples of 30-60-90 criteria for different roles
Customer service
In the first 30 days, the focus can be on understanding products, policies, fulfillment system, brand language and escalation rules. Readiness appears when the person can correctly classify common situations, locate applicable guidance, and conduct parts of the care with follow-up.
Up to 60 days, delivery can evolve into simple or moderate services with sample review. The criteria should not just be number of tickets, but adherence to procedure, clarity of communication, correct recording and ability to recognize when the case needs to be escalated.
Close to 90 days, autonomy may mean conducting the routine cases expected for the position, maintaining reliable records and asking for support in real exceptions. The manager remains available, but stops reviewing each step.
Administrative or financial analyst
In the first month, the person needs to understand the calendar, controls, data sources, responsible parties and risk points. An initial piece of evidence is being able to explain the cycle and execute a low-criticality part with an upcoming conference.
In the second month, you can undertake a report, conciliation or defined routine with review by the person responsible. Readiness depends on precision, respect for deadlines, traceability of information and the ability to identify discrepancies before completing the work.
In the third month, the criteria may be to conduct the routine cycle with review only in exceptions, communicate deviations in advance and propose corrections when an inconsistency is identified. This demonstrates operational mastery without requiring the person to already know all the extraordinary situations.
Software development
In the first 30 days, the goal should be to make the environment, workflow, and team standards understandable. Instead of demanding a large deliverable, it makes more sense to see if the person can prepare the environment, navigate the codebase, run tests, understand the review process, and contribute to a low-risk change with support.
Up to 60 days, the person can take on a small end-to-end task within a known component, with normal team review. Criteria include code quality, testing, communication about roadblocks, and proper use of the review process, not just speed to close the task.
Up to 90 days, autonomy can increase for a delimited set of demands, as long as the person recognizes architectural, security or product impacts that require discussion. The goal is not to eliminate technical review, which is part of the job, but to reduce the need for guidance on each basic decision.
Consultative sales
In the first month, the person needs to understand the offer, customer profile, funnel stages, CRM and qualification criteria. You can attend meetings, simulate conversations and demonstrate that you know how to record opportunities and explain the value proposition without improvising undue promises.
In the second month, start conducting discovery parts, demonstrations or follow-ups with feedback from the manager. Readiness can be observed in the quality of the questions, the recording of information and the ability to advance or disqualify an opportunity according to defined criteria.
In the third month, the salesperson can take on opportunities compatible with their level, keeping the process and forecast reliable. The company can start applying progressive commercial goals, but it must make it clear from the beginning when and how these goals are included in onboarding, preventing charges from appearing only at the end of the period.
How to assemble the document without turning it into bureaucracy
A useful plan can fit into a few pages. For each phase, record what needs to be learned, what actual deliverable will be used to practice, what evidence will show readiness, who is responsible for supporting, what level of supervision will be maintained, and when check-in will happen. These elements are enough to create an objective conversation without setting up a parallel evaluation system.
The plan must be reviewed at the beginning with the new employee and updated when reality shows that some assumption was wrong. Adjusting does not mean silently lowering the standard. It means preserving the expected result and modifying the path when
training, priorities or working conditions change.
It is also worth keeping the master model separate from the plan for each function. The model organizes questions and responsibilities; the specific plan defines criteria. This avoids the mistake of copying the same expectations for an analyst, a salesperson and a developer just because they all went through the same institutional integration.
The best sign of a well-designed onboarding is the absence of surprises
When the 90-day check-in arrives, no essential questions should be new. The person already knows what is expected, what evidence counts, where they still need help and what autonomy is being built. The manager can explain the decision with observable examples, and HR can verify whether the process was executed consistently.
The calendar remains useful, but as a cadence. What determines advancement is the demonstrated ability to assume the next level of responsibility. It's this combination of clear goals, real support, observable evidence, and frequent feedback that turns the 30-60-90 plan into true integration rather than an assessment hidden behind three dates.
