First 90 days as a data analyst: a probation-period plan
In this article (7 sections)
Related programme: the Data Analytics with Generative AI course (3–4 months; ₹90,000 (inclusive of GST) · EMI from ₹4,000/month).
A good first 90 days as a data analyst has three stages: learn how the business and its data actually work (weeks 1–4), deliver one or two small, visible pieces of work that someone uses (weeks 5–8), and then own a recurring task and prepare evidence for your probation review (weeks 9–13). Agree the plan with your manager in the first week so that "confirmed" has a definition you both know.
Probation in India is commonly three or six months, as stated in your appointment letter. This plan covers the first 90 days; on a six-month probation, the last stage simply runs longer.
Week 1: agree what success means
In your first one-to-one, ask four questions and write the answers down:
1. What will make you confident to confirm me at the end of probation?
2. Which recurring reports or dashboards will I own, and by when?
3. Who are the three people outside this team whose questions I will answer most?
4. How do you want to hear about problems: message, email, or in our one-to-one?Send the answers back as a short email: "Summarising what we agreed today." That email becomes the reference for your review.
The plan, week by week
| Weeks | Focus | Concrete output |
|---|---|---|
| 1–2 | Access, tools, people | Access to warehouse, BI tool and repo; a list of the ten most used tables and their owners |
| 2–4 | How the business measures itself | A one-page glossary: how "active customer", "order" and "revenue" are defined here, with the SQL for each |
| 3–4 | Shadow the recurring work | Run last month's main report yourself and match the published numbers |
| 5–6 | First small win | Fix or automate one thing that annoys the team: a manual Excel step, a broken filter |
| 7–8 | First answered question | Answer one stakeholder question end to end, with the query, the result and the caveat |
| 9–11 | Ownership | Take over one recurring report; publish it on time twice without help |
| 12–13 | Probation evidence | Write the review document below |
The glossary is the most useful thing you will write
Every company defines its metrics slightly differently, and most errors by new analysts come from assuming a definition. Build a glossary as you go:
Metric Definition here Source table Owner
Active customer Placed an order in the last 90 days orders (status=paid) Growth team
Revenue Paid order value, excl. GST, excl. orders, refunds Finance
refunds within 30 days
New customer First paid order ever, by phone number customers CRM teamWhen a number you produce differs from someone else's, the glossary usually explains why. It is also the document your manager is most likely to share with the next new joiner, which is a quiet way to be noticed.
Matching last month's report
In weeks 3–4, pick the main recurring report and reproduce it from raw tables. If your numbers match, you understand the pipeline. If they do not, you have found either a gap in your understanding or an error in the report. Raise the second carefully:
Hi Sunil, while reproducing the August sales report I get ₹4,12,30,000 against the
published ₹4,18,90,000. The difference is orders cancelled after dispatch; the
report includes them. Is that intended? Happy to walk through the query.The figures are illustrative. The tone matters: a question, not an accusation.
Habits that protect you in probation
- Write down every request with the date and who asked. When two people ask for conflicting things, you can show both to your manager.
- Share a caveat with every number. "This excludes returns processed after the 5th."
- Save your queries in the team repo or a shared folder, with a comment on what each answers.
- Ask for feedback at day 45, not day 90. It gives you time to act on it.
The probation review document
Bring a one-page summary to the review. A fictional example:
Probation summary - Kavya Nair, Data Analyst - 1 Aug to 31 Oct
Agreed in week 1: own the weekly sales dashboard; support Category team queries.
Delivered
- Weekly sales dashboard: owned since week 9; published on time 5 of 5 weeks.
- Automated the returns reconciliation; manual time went from about 3 hours to
20 minutes a week (team estimate).
- Metric glossary: 14 metrics defined, now linked from the team wiki.
- Answered 11 Category team requests; list with dates attached.
Learnt
- Warehouse schema for orders, returns, inventory; dbt model structure.
Next quarter
- Take over the monthly board pack; learn the forecasting model the team uses.Every line points to something your manager can check. That is what makes confirmation straightforward.
If probation is extended
Ask exactly what needs to change, by when, and how it will be measured, and confirm it in writing. Treat it as a second, shorter version of this plan.
If you are still at the stage of getting the first role, the Data Analyst programme spends its 3–4 months on the SQL, Excel and dashboard work that fills weeks 1–13 above, and a free skill check will show where you stand now.
Pankit Kumar has 10 years in Data Science & AI, building and shipping production systems in regulated pharma and clinical environments. He is a freelance trainer at Boston Institute of Analytics, AnalytixLabs and Scaler, and has taught this material to thousands of working professionals.
This article is part of our Data Analytics with Generative AI programme — 3–4 months. The full analyst stack — Excel, SQL, Power BI and Python pipelines — then a generative-AI layer you can prove is right.
Explore Data Analytics with Generative AI