How to Get Promoted from Data Scientist IC3 to IC4 at Meta
Building your promotion case?
Get AI coaching on what to document and how to present it.
You're a few months into your first data science role at Meta. Your SQL is clean, your dashboards are accurate, and your manager seems happy enough with your work. But you've seen the Blind posts about people getting exited at the 24-month mark, and "seems happy" doesn't feel like the same thing as "ready to promote you."
It probably isn't. IC3 at Meta is not a terminal level. There's an explicit expectation that you advance to IC4, and a clock running in the background. The IC3-to-IC4 promotion is the most straightforward level jump for data scientists at Meta, but straightforward doesn't mean automatic. You still need a Performance Summary Cycle (PSC) packet that survives calibration, a manager who builds the case, and evidence you've already been working like an IC4.
What Changes from IC3 to IC4
IC3 is Meta's entry-level data science role. IC4 is where Meta considers you a fully independent data scientist. The title shifts from Associate Data Scientist to Data Scientist, and the expectations change in specific ways.
| Dimension | IC3 (Associate Data Scientist) | IC4 (Data Scientist) |
|---|---|---|
| Autonomy | Works on well-defined tasks with guidance from senior DS | Independently scopes and executes analyses end-to-end |
| Scope | Individual tasks and feature-level analyses | Owns analyses tied to product team roadmaps |
| Product sense | Learns the team's business logic and data sources | Identifies opportunities and shapes what the team focuses on |
| Methodology | Executes assigned experiments and analyses | Designs experiments, develops KPIs, monitors A/B tests independently |
| Communication | Asks good questions, absorbs feedback | Communicates findings to product and engineering partners proactively |
| People | Minimal expectation | Begins helping onboard new teammates and answering methodology questions |
The core shift: at IC3, someone gives you a question and tells you how to answer it. At IC4, you identify the right question, figure out the approach, run the analysis, and deliver the answer to stakeholders who will act on it.
How Meta Promotions Work at IC3-IC4
Meta's promotion process is manager-driven with calibration oversight. Your manager builds your promotion case and presents it during calibration alongside your PSC results. The calibration committee normalizes ratings and promotion decisions across the data science org.
The process follows a clear sequence:
- You write your self-review (~1,000 words) during the PSC window
- Peer feedback is collected from 3-5 nominators you choose
- Your manager drafts their assessment and proposes your rating
- Calibration happens: managers argue for their reports' ratings and promotions against competing claims
- Results come through the review tool, followed by a 1:1 with your manager
Your PSC is evaluated across Meta's performance dimensions, adapted for data science:
- Analytical Impact: what analyses you delivered and what product decisions changed because of them
- Methodology: rigor of your statistical approach, experiment design, data quality standards
- Direction: how you influenced what the team worked on, not just executed what was assigned
- People: mentorship, collaboration, raising the team's analytical quality bar
At IC3, Analytical Impact and Methodology carry the most weight. The committee wants evidence that you can deliver rigorous, independent analyses that move product decisions. Direction and People matter less at this transition than they will for IC4 to IC5, but showing early signal strengthens your case.
The rating that matters: an Exceeds (EE) rating signals promotion readiness. Greatly Exceeds (GE) makes the case very strong. Meets All (MA), the most common rating at roughly 45% of the org, means you're performing solidly at IC3 but aren't generating promotion momentum.
The backdrop: Meta has an up-or-out policy at IC3. You have roughly 24 months to make IC4. Miss that window and you'll be exited. Two consecutive Meets Most (MM) ratings trigger an automatic PIP even before the deadline arrives. Every PSC cycle counts.
How Long IC3 to IC4 Should Take
| Pace | Timeline | What's happening |
|---|---|---|
| Fast | 12-18 months | Strong from the start; may have been under-leveled at hire; IC4-scope analyses early |
| Standard | 18-24 months | Solid contributor, steady growth across two PSC cycles |
| Slow (danger zone) | 24+ months | Hitting the up-or-out boundary; something structural needs to change immediately |
The compensation jump matters here. Based on Levels.fyi data, median total comp at IC3 is roughly $170K compared to approximately $273K at IC4. That's a $100K+ increase, driven primarily by larger stock grants. Every cycle you wait costs real money, and the up-or-out deadline means waiting too long costs you the job.
What Actually Gets You Promoted
Own an analysis end-to-end
The clearest signal of IC4 readiness is completing an analysis from problem identification through stakeholder delivery without someone else defining every step. This doesn't need to be a massive project. It could be an experiment design for a product launch, a deep-dive into why a metric dropped, or a forecast model that shapes next quarter's goals. What matters is that you drove it: you identified the question, designed the approach, ran the analysis, and delivered findings that changed a product decision.
If your current workload is entirely tasks your manager or senior DS scoped for you, ask for work where you own the full analytical cycle.
Build product sense early
IC3 data scientists who stay in SQL-and-dashboard mode get stuck. IC4 means understanding why the product team cares about a metric, not just how to calculate it. Read the product roadmap. Sit in on product reviews. When you present analysis results, connect them to what the team is deciding, not just what the data shows.
The data scientists who advance fastest are the ones who start saying "based on this data, I think we should..." rather than "here are the numbers you asked for."
Write self-reviews that argue your case
Your self-review is one of the primary documents the calibration committee reads. For each contribution, cover what you did, what impact it had on product decisions, and how it went beyond IC3 expectations.
"I designed and ran the A/B test for the new notification frequency model, identified that the 20% reduction in push notifications increased 7-day retention by 3%, and presented findings to the product team. They shipped the change to 100% of users based on my analysis."
That's an IC4 self-review paragraph. Compare it to: "I ran several A/B tests and built dashboards for the notifications team." Same work, completely different calibration signal.
Make your work visible
Meta's internal tools, including Workplace, serve as evidence in your PSC. Data scientists who share analysis summaries, post methodology write-ups, or answer analytical questions publicly create a trail of artifacts that feed into peer feedback and self-review content.
This isn't about self-promotion for its own sake. It's about making your work visible to people who will later write your peer feedback and to the manager who needs material to defend your rating in calibration.
Have the promotion conversation early
After your first full PSC cycle, ask your manager directly: "What does IC4 readiness look like for me? What evidence would make my case clear in calibration?"
This gives you a target. Your manager is the person who presents your case in calibration. They need to know you're aiming for IC4, and they need specific wins to build that case around.
Mistakes That Keep Data Scientists at IC3
Staying in dashboard mode. The most common pattern among IC3 data scientists who stall is building dashboards and running queries without ever stepping into analysis that shapes product direction. IC4 requires going from "here's the data" to "here's what the data means and what we should do about it." If every piece of work you've done was a reactive request, your manager doesn't have the material to argue for a level change.
Not tracking your wins. When the PSC window opens, you have a week or two to recall months of work. Data scientists who log their analyses and their impact as they happen write stronger self-reviews. Those who scramble to remember what they did six months ago produce vague summaries the committee can't evaluate.
Treating data science as a service function. If product managers hand you questions and you hand back answers, you're operating as a support function. IC4 data scientists identify questions the product team didn't think to ask. They look at a metric trend and flag a problem before anyone notices. The shift from reactive to proactive is what calibration evaluates.
Ignoring the People dimension. IC3 data scientists sometimes go heads-down on their analyses for 12 months without helping anyone else. At IC4, even early signal matters: answering methodology questions from teammates, helping a new hire understand the data model, or reviewing someone else's experiment design. Zero people signal is a gap calibration notices.
Assuming your manager tracks everything. Managers at Meta can have 8-12 reports. They're not remembering every analysis you ran or every experiment you designed. If you're not surfacing your contributions in 1:1s and internal posts, your manager's calibration packet will reflect what they remember, not what you actually did.
Frequently Asked Questions
How long does it take to get promoted from IC3 to IC4 at Meta?
Most data scientists who get promoted spend 18-24 months at IC3. Strong performers with good project fit can do it in 12-18 months. The minimum is typically one full PSC cycle of strong results, though two cycles is more common. Meta's up-or-out policy sets a deadline of roughly 24 months. Data scientists who haven't been promoted by then are exited.
What's the pay difference between IC3 and IC4 at Meta?
Based on Levels.fyi, median total compensation jumps from roughly $170K at IC3 to approximately $273K at IC4. That's a $100K+ increase, driven by base salary increases and significantly larger stock grants. This is one of the largest percentage jumps in Meta's data science ladder.
Is the IC3 to IC4 promotion considered easy?
Relative to later promotions, yes. The bar for IC3 to IC4 is lower than IC4 to IC5, and the expectations are more concrete: run analyses independently, design experiments without hand-holding, communicate findings clearly to stakeholders. But it's still a calibration decision that requires your manager to present evidence. "Easy" doesn't mean it happens without effort.
What PSC rating do I need for promotion from IC3 to IC4?
An Exceeds (EE) rating is the standard promotion-track signal. Greatly Exceeds (GE) makes the case even stronger. Meets All (MA), the most common rating at roughly 45% of the org, means you're performing well at IC3 but doesn't generate promotion momentum on its own. One or two cycles of EE with clear IC4-scope evidence is the typical profile.
Should I switch teams if I'm stuck at IC3?
Only if the problem is your team, not your skills. If your team doesn't have IC4-scope analytical work (rare, but possible on teams with limited experimentation), switching can help. But a team change resets your context, and given the 24-month clock, switching mid-tenure is risky. If your manager is the issue, try having the direct conversation first. If nothing changes after a full PSC cycle, then move quickly.
CareerClimb tracks your analytical wins across Meta's performance dimensions as they happen, maps them to what calibration evaluates, and shows you exactly what evidence you're missing before your manager presents your case. When the next PSC window opens, your self-review is backed by documented impact, not memory. Download CareerClimb