Cold email a hiring manager for a data role: a draft you can adapt

Find the manager, not just any employee

The person you want is usually a data science manager, analytics lead or head of data for the team that posted the role. Three ways to identify them:

• The job posting sometimes names the team ("Risk Analytics", "Growth Data Science"). Search LinkedIn for that team name plus the company and titles like "Manager", "Lead" or "Head".
• Company engineering or data blogs often have author names. An author of a post about the team's work is likely close to the hiring decision.
• Conference or meetup talk listings name speakers and their teams.

Do not guess email addresses by trial and error and send to all variations. If you cannot find a published address, send the same short message on LinkedIn instead.

Pick the one piece of evidence

A hiring manager decides in seconds whether you are relevant. Choose one project whose problem resembles theirs. A fictional example:

If none of your work is close, the email will not help much. Build the relevant project first, or apply to teams whose problems you have worked on.

The draft

Subject: Data Scientist (Job ID 4521): churn modelling work

Hi Ananya,

I applied for the Data Scientist role on your retention team this week. I am
writing because the posting mentions predicting churn timing, and that is the
problem I spent my last project on.

I modelled churn on a synthetic telecom dataset as a survival problem rather than a
yes/no classifier, and wrote up why the classifier gave a misleading picture of
when customers leave: [link to write-up].

If it is useful, I would be glad to walk you through it in 15 minutes. Either way,
thank you for reading.

Rahul Verma
[phone] | [LinkedIn URL]

Line by line:

• Subject includes the job ID so it can be matched to the application, and a topic that is relevant to them.
• First line says you applied properly. Managers dislike being used to skip the process.
• Second paragraph is the whole pitch: their problem, your approach, one link. Label synthetic data as synthetic.
• The ask is small and optional. "Can you get me an interview" is a large ask from a stranger; "15 minutes if useful" is not.
• No attachment. Unknown attachments get filtered or ignored; the link and your application are enough.

A version for when there is no open role

Some teams hire without posting widely. You can still write, but the ask changes from the role to advice or future openings:

Subject: Question about the forecasting work in your team's blog post

Hi Ananya,

I read your post on moving store-level forecasts to hierarchical reconciliation. I
tried the same idea on a synthetic grocery dataset and found reconciliation helped
low-volume stores most: [link].

I am looking for data science roles in forecasting. If your team expects to hire
later this year, I would be grateful to be considered, and I would welcome any
feedback on the write-up.

Rahul Verma

This works only if the reference is genuine. A manager can tell when a sentence about their work was written without reading it.

Timing and follow-up

• Send on a working day morning. There is no reliable evidence for a magic hour, so do not overthink it.
• Follow up once after about a week if there is no reply, in the same thread, with one new line, such as a short update to the project.
• Stop after that. Two unanswered emails is a clear answer.

Track what you sent

Log each email with the date, manager, team, job ID, the project you linked and whether they replied. After ten or fifteen emails, look at which projects got replies; that tells you which work is persuasive and which needs improving. The job application tracker has a layout you can reuse, and the referral message drafts cover the warm-contact route.

A cold email is only as strong as the project it links. The Data Science programme is organised around evaluated, written-up projects of this kind, which are the pieces worth putting in the second paragraph.