Data Analyst
Citation is building a workspace for people who turn research into something worth reading. We bring sources, notes, drafts and references together so a writer can follow an idea from the material that inspired it to the document they share. Our work serves the everyday decisions behind a paper, an article, a report or a longer research project.
We care about the connections between those decisions. Finding a source is only useful when its details can be reviewed. An AI suggestion needs to fit the writer’s intention and remain theirs to accept or reject. A finished document should carry its references with it. We want the product to make those steps easier to understand while leaving room for the judgment that good writing requires.
As a startup, we value people who can follow a problem across disciplines and stay close to the person using the result. A clear explanation, a dependable interaction and a thoughtful technical decision all contribute to the same experience. Joining Citation means helping shape both the product and the way we build it, with care for the details and a willingness to learn from the work.
The opportunity
We are looking for a Data Analyst who wants to understand the work behind a number. Citation contains connected activities: finding a source, beginning a document, revising a passage, reviewing evidence and returning to an unfinished draft. A useful analysis needs to distinguish those activities and explain what they reveal about the person’s experience. Your role is to help the team ask better questions and make decisions supported by trustworthy definitions and evidence.
The work begins before a dashboard. We need to agree on what an event means, how a journey is represented and which limitations affect a comparison. A change in a count may reflect a product improvement, a different audience, an instrumentation change or an unreliable denominator. We want someone who can investigate those possibilities and explain the result clearly to people who are making product, growth and operational decisions.
This role offers the chance to develop an analytical foundation close to the product. You will work with the people implementing events, interpreting user feedback and choosing what to build next. The aim is a practical system of definitions, analyses and reusable questions that helps Citation learn without collecting more information than the work requires.
What you’ll work on
- Define meaningful measures for important product journeys, including activation, task completion, repeat use and friction. Explain the unit of analysis, denominator, time window and exclusions so a metric can be understood and reused consistently.
- Audit instrumentation with engineering and product contributors. Check event meaning, duplication, missing data and state transitions, and verify that a change in the interface has not silently changed the interpretation of the resulting data.
- Investigate product questions using appropriate queries, cohorts and comparisons. Follow an unexpected result through possible explanations and distinguish a descriptive pattern from evidence that supports a stronger causal conclusion.
- Build concise reports and analytical views that support decisions. Include enough context to explain what changed, what is uncertain and which next action the evidence can reasonably inform.
- Work with growth contributors on acquisition and continued use. Keep channel definitions and attribution limitations visible, and help compare audiences without allowing differences in exposure or timing to disappear inside an aggregate.
- Support experiments and release reviews with clear hypotheses and practical measurement plans. Identify the observations needed before a change ships, and review results with attention to sample size, selection and unintended effects.
- Maintain useful analytical documentation and careful data handling. Keep definitions and transformations discoverable, limit unnecessary sensitive content and make it easier for another person to reproduce an important result.
What you’ll bring
- You can use SQL to investigate product or business questions and explain how the query reflects the definition being measured. You are comfortable checking joins, event duplication, time boundaries and the effect of missing data.
- You have produced an analysis that influenced a decision and can describe both the finding and its limitations. We want to understand how you moved from an ambiguous question to evidence someone could use.
- You understand cohorts, funnels, rates and comparisons well enough to challenge a convenient but misleading metric. You keep the denominator and observation window attached to the result.
- You can communicate clearly with people who do not work in analysis every day. You can choose a useful table or visualization and explain the practical meaning without hiding uncertainty behind technical vocabulary.
- You are comfortable collaborating on instrumentation and data quality. You can investigate a discrepancy with engineering and help decide whether the issue is in the product behavior, collection process or analytical definition.
- You bring judgment to the amount of analysis a decision needs. You can distinguish a quick directional check from a result that requires more careful validation and explain the difference to the team.
Useful, but not required
- Experience with software product analytics, event based data or experimentation would be useful when connecting a product interaction to a longer writing journey.
- Familiarity with Python, reproducible notebooks or analytical transformation tools would help with investigations that extend beyond a single query or dashboard.
- Experience with privacy conscious measurement, retention analysis or qualitative research would add useful perspective when deciding what the data can explain and what requires another kind of evidence.
Your first few months
- Learn the product and its important journeys, inspect the current event definitions and reproduce a small set of existing measures. Identify the most consequential data quality questions and agree on one analysis that can inform an immediate product decision.
- Complete that analysis with a clear definition, reproducible method and useful explanation. Work with engineering on any collection issue it exposes, and document the assumptions so the result can be revisited as the product changes.
- Establish a compact set of dependable measures for an area of the product. Help the team interpret changes, support a release or experiment review and propose the next analytical improvement based on the decisions people need to make.
Who you’ll work with
You will work with engineering on data collection, product on behavior and outcomes, growth on audience journeys and customer experience on the context behind recurring problems. Quantitative and qualitative evidence should inform each other.
We value an analyst who can ask a precise question and make a useful challenge. If a result does not support the conclusion being drawn, explain why and suggest what evidence would help. Keep the discussion connected to the decision rather than only to the correctness of a chart.
What good work looks like
- Important product measures have stable, understandable definitions, and the team can explain what a change in those measures does and does not mean.
- Analyses lead to clearer decisions because their methods, limitations and practical implications are visible and reproducible.
- Instrumentation and data quality improve alongside the product, making future questions easier to answer without unnecessary collection or repeated manual reconstruction.
How we work
We start with a concrete user problem and choose a focused piece of work that can make it better. Important decisions should be understandable in writing, with the relevant context, tradeoffs and open questions available to the people contributing. We share unfinished work when feedback can still change the outcome and protect enough time to think, investigate and build.
Ownership includes asking for help, explaining a constraint and following through after the first version is delivered. We expect people to care about the whole experience and to challenge an assumption with evidence and consideration for the person who made it. Clear communication and dependable follow through matter more than how visible someone can make their activity.
Each working arrangement is discussed individually. We clarify responsibilities, location, availability, compensation and the practical support needed for the role before agreeing to start. You do not need to match every optional experience listed here. Show us the work you can do, the judgment behind it and the areas you are ready to learn.
Getting to know each other
1. Introduce yourself
Tell us which area interests you, how you prefer to work and where we can see a relevant example. Include your location, time zone and availability.
2. Discuss the fit
We review the introduction against the work we need to do. When there is a relevant opportunity, we arrange a conversation about your experience and the problem.
3. Explore the work
Walk through a project and explain your decisions. If a practical exercise would help, agree on its scope and expectations before it starts.
4. Agree on the details
Clarify responsibilities, working arrangement, compensation and start date together. An introduction does not require you to share confidential work or customer data.
Apply for this role
Share an analysis that changed a decision or corrected an assumption. Explain the original question, data, definitions, method and limitations, along with your own contribution. Include a query, notebook, report or anonymized walkthrough where possible. We are especially interested in a case where careful interpretation mattered more than a complicated technique.