You've been staring at job postings for three weeks. You've tailored your resume. You've done the LeetCode problems. You've even built a portfolio project analyzing something you actually care about. And yet the applications keep going into the void — no responses, no feedback, just the algorithmic silence of an ATS that doesn't know you exist.
Here's the thing most job search advice skips: the people who land data roles fastest aren't necessarily the ones with the most skills. They're the ones who figured out how to make human connections before a job posting ever appeared. Informational interviews — genuine, low-stakes conversations with working data professionals — are the single highest-leverage activity available to a job seeker who doesn't already have a warm network. Done well, they get you insider knowledge about what companies actually need, they build relationships that surface unadvertised opportunities, and they give you a loop of feedback that no tutorial can provide. Done poorly, they waste everyone's time and leave you right back at that job board.
This lesson is about doing them well. By the end, you'll know exactly how to identify the right people to contact, craft a message that gets a response, conduct a conversation that's genuinely useful for both parties, and — critically — what to do afterward so the relationship actually goes somewhere. This isn't networking-as-performance. It's professional relationship building with a strategy behind it.
What you'll learn:
You don't need to be a seasoned networker. But you should have:
If you're still in the exploratory phase of figuring out which data role you want, informational interviews are actually a great tool for that. Just be honest with yourself and the people you contact that you're still exploring — that changes the framing of your outreach slightly, and we'll cover how.
Before we get into tactics, let's be clear on the underlying mechanism. Informational interviews work because of a well-documented psychological principle: people like to help, and people like to talk about work they find meaningful. When you ask someone a genuine question about their career and then listen attentively, they feel useful and respected. That creates goodwill. Goodwill becomes referrals, tips, and remembered names when a position opens up.
What kills this dynamic immediately is treating the informational interview as a covert job application. The person on the other end of that conversation has been in your position. They know what you're doing. If your "casual conversation" is actually a stealth pitch for a job, they feel manipulated — and you've burned the relationship before it started.
The second most common failure mode is the opposite problem: people are so anxious about not seeming like they want a job that they make the conversation completely aimless. You spend 30 minutes asking vague questions about "what's a day in your life like" and learn nothing useful. The person goes back to their work and doesn't think about you again.
The goal is a genuine conversation with a clear intellectual focus. You're not hiding that you're job searching — that would be weird and dishonest. You're also not asking for a job. You're asking for information, perspective, and insight. That's a thing people can give freely. A job offer is a thing with real stakes attached. The distinction is real, not just semantic.
The core reframe: An informational interview is not a networking tactic with a job at the end of it. It's the first conversation of a professional relationship that might, over time, lead to an opportunity — or might just make you smarter and better connected. Both outcomes are worth having.
Most job seekers approach outreach as a numbers game: message 50 people, hope 5 respond. This is both inefficient and demoralizing. A better approach is targeted outreach where you contact fewer people but with much higher relevance and specificity.
Your outreach list should come from a clear set of criteria based on your actual goals. For someone targeting a data analyst role at a mid-sized tech company, the ideal contacts look something like this:
This doesn't mean you only talk to people who are perfect analogs to your ideal job. Sometimes the most useful conversations are with people two or three years further down the path than you, or people who left a company you're targeting and can give you honest inside information.
LinkedIn is the primary channel, but use it intelligently. The search filters are more powerful than most people realize. You can search by job title, company, geography, and mutual connections. Filter specifically for people who list tools you're learning in their "Skills" section — that's a proxy for technical alignment.
The Alumni filter is underused and extremely valuable. If you attended any university, you can filter your LinkedIn network by alumni and then cross-reference with job titles. People are statistically more likely to respond to outreach from fellow alumni — use this every time.
Twitter/X and Mastodon (for the data community specifically) are excellent. Data professionals who write publicly about their work — tweeting about dbt releases, posting about their SQL optimization wins, writing threads about analytics engineering — are already in a mode of sharing knowledge. They're easier to contact because they've already demonstrated a willingness to educate. Engage with their content genuinely before cold-messaging them.
Data community Slack groups like the dbt Slack, the Locally Optimistic Slack (analytics-focused), and the MLOps Community Slack are direct access to working professionals. These are not job boards disguised as communities — they're places where people actually talk shop. Participate in channels, answer questions where you can, ask smart questions in public. Relationships built here lead naturally to DM conversations that feel warm, not cold.
Substack and Medium — find data professionals who write about their work. If someone wrote a piece about how they built a data quality framework and you found it useful, that's an extraordinarily good entry point for outreach. You're not cold-messaging them; you're responding to something they already put into the world.
Target 3-5 high-quality outreach messages per week rather than blasting 50 generic ones. Your response rate on a thoughtful, specific message is dramatically higher than on a template — often 30-50% versus single digits. More importantly, the conversations you get will actually be useful.
This is where most people spend too little time. The message is the entire thing. If the message is weak, nothing else matters.
A good informational interview request has five components, and they all need to be present:
Let's look at what each of these looks like in practice.
The hook is what separates your message from the 15 other "I'd love to connect" messages in this person's inbox. It needs to demonstrate that you actually know who they are and why you're reaching out to them specifically.
Bad hook: "I came across your profile and was impressed by your experience in data."
This is useless. It could apply to anyone with a data job. It signals that you searched for "data analyst" and messaged the first ten results.
Good hook: "I read your post about how your team rebuilt their attribution model in dbt last year — specifically the part about handling multi-touch attribution with incremental models. I've been working through a similar problem in a side project and your approach gave me a new way to think about it."
This is specific, it demonstrates real engagement, and it tells the person exactly why you reached out to them rather than someone else. Even if the person posts rarely, you can often find something — a comment they left on another post, a job change, a conference talk listed on their profile.
If you genuinely cannot find a hook, you can fall back to company specificity: "I've been following [Company]'s work on [specific product/initiative] and I know from reading job descriptions that your data team is working on [X]." This is less personal but more specific than the generic compliment.
Here's a complete example, with each component annotated:
Subject line: Data question from an analytics engineering student — 20 min?
Hi Priya,
[Hook] I came across your piece on the Locally Optimistic blog about building a metrics layer without a semantic layer tool — the part about using dbt exposures to document metric definitions really clicked for me after I'd been struggling to explain that pattern to my study group.
[Who you are, briefly] I'm transitioning into data engineering from a finance background — I spent three years doing FP&A at a mid-size SaaS company, which means I can write SQL in my sleep and understand the business side, but I'm still building my engineering credibility.
[The specific ask] I'd love to ask you about how your team thinks about the boundary between analytics engineering and data engineering — specifically when that line gets blurry and how you navigate it. It's a question I keep running into as I target analytics engineering roles.
[Time commitment + low-friction path] Would you be open to a 20-minute call in the next few weeks? I'm flexible on timing and happy to work around your schedule. Even a few thoughts by email would be incredibly useful if a call doesn't work.
Thanks for considering it — [Your name] [LinkedIn URL or brief credential]
Let's dissect what this message is doing right:
When you're still exploring which data role to pursue:
The framing shifts from "I'm targeting X" to "I'm trying to understand the landscape." Be explicit about this:
"I'm at a point in my learning where I'm genuinely trying to figure out whether analytics engineering or data science is a better fit for where I want to go — and I'm curious how people in the field think about those paths."
This is honest and actually more interesting to talk about than someone who's narrowed in. Just make sure your questions reflect genuine uncertainty, not just a covering statement over an implicit pitch.
When you have a mutual connection:
Lead with the connection immediately:
"Hi Marcus — our mutual connection Aisha Osei suggested I reach out to you. She thought you'd have an interesting perspective on..."
A mutual connection reference dramatically increases response rates and changes the relational dynamic from cold to warm. Always use this when available, and always be specific about what Aisha said — "she thought you'd be good to talk to" is weaker than "she mentioned you went through a similar career pivot."
When you're reaching out to someone very senior (VP, Director, Principal):
Senior people get more outreach and have less time. Your message needs to be shorter, your hook needs to be sharper, and your ask should be even more specific. Two short paragraphs maximum. Make your question so specific that they can see exactly what 20 minutes with you looks like — and see that it won't turn into a 90-minute obligation.
When you're reaching out via Twitter/Slack instead of LinkedIn:
The tone is slightly more casual but the structure is the same. On Twitter, you might reference a specific tweet. In a Slack DM, you might reference a conversation in a public channel. The hook still needs to be specific. The ask still needs to be clear. The time commitment still needs to be stated.
Most people treat outreach timing as an afterthought. In reality, timing affects response rates significantly, and it also governs how the relationship progresses.
For LinkedIn and email, Tuesday through Thursday, between 9 AM and 11 AM in the recipient's time zone is the sweet spot backed by email marketing data. Monday mornings are cluttered with catch-up tasks. Friday afternoons are mentally checked out. Wednesday mid-morning is often the single best moment — people are in the rhythm of the week but haven't yet hit the productivity crunch of Thursday.
This matters more when you're emailing someone directly than when using LinkedIn messages — LinkedIn messages don't sit in an inbox the same way — but the principle holds.
Avoid major industry events — if there's a big data conference (Coalesce for the dbt community, Data + AI Summit for Databricks, etc.), the people you want to talk to are either at the conference or swamped with preparation and follow-up. Wait a week on either side.
Pay attention to company news. If a company just announced layoffs, a major acquisition, or a significant product failure, the data team is likely in crisis mode. Send your note two or three weeks later when the dust has settled. Conversely, a company that just announced a big funding round or a major product launch has a data team that's probably energized and busy — acknowledge the news in your hook, but understand they're heads-down.
If you don't get a response in a week, one follow-up is appropriate. More than one is not.
Follow-up message:
Hi Priya — just wanted to make sure this didn't get buried. I know inboxes can be relentless. If now isn't a good time, no worries at all — I appreciate you considering it either way.
That's it. Short, no guilt-tripping, no "just checking in" (which signals you're tracking them), and an explicit release valve so they don't feel obligated. If they don't respond after the follow-up, let it go. Follow up again in three months if something specific and relevant comes up — not to rehash the original ask, but to share something (an article, a project you finished) that's genuinely relevant to them.
The 48-hour exception: If someone follows you back on Twitter, engages with your content, or interacts with your comment on their post within 48 hours of your initial outreach, that's a signal of interest. You can follow up sooner — but do it by building on the interaction, not just resending your original ask.
Don't send all your outreach on the same day. Stagger it. Three reasons:
A sustainable cadence is 3-4 outreach messages per week, with space to actually have the conversations and follow up properly.
You got a yes. Now what?
When someone agrees to a call, respond quickly (within 24 hours) and make the scheduling easy. Offer specific time slots in their time zone:
"That's great — thank you so much. I'm flexible on timing. Would any of these work for you? Tuesday the 14th at 10 AM PT, Thursday the 16th at 2 PM PT, or Friday the 17th at 11 AM PT? Happy to find other slots too."
Offering a scheduling tool like Calendly is appropriate if you use one, but some people find it impersonal. Gauge from the tone of their reply — if they're warm and conversational, offer times personally. If they're brief and efficient, a Calendly link is probably fine.
Confirm the format: video or phone? Most data professionals are comfortable with both. Video is generally better for building a real connection, but don't insist on it — some people prefer phone because they can pace or multitask less conspicuously.
Do actual research. Spend 20-30 minutes before the call so you arrive as a peer in conversation, not as someone reading from a blank page.
Specifically look for:
This research serves two functions: it generates better questions, and it signals respect when you demonstrate in the conversation that you actually know who they are.
Prepare 8-10 questions, knowing you'll only get through 4-6. Having more than you need means you'll never run out of things to ask, and you can adapt in real time based on where the conversation goes.
Good questions are open-ended, specific, and invite reflection — not yes/no, not easily Googleable, not ones that require them to reveal proprietary information.
Career trajectory questions (use 1-2):
Role and team questions (use 2-3):
Industry/landscape questions (use 1-2):
Advice questions — use sparingly, at the end:
Never ask: "Can you help me get a job?" or "Can you refer me to the hiring manager?" This is the question that poisons informational interviews. If the relationship develops and the person wants to make a referral, they will. Asking for it directly in a first conversation violates the implicit contract of the informational interview and makes the person feel used.
The meeting is live. Here's how to manage it in real time.
Start by thanking them briefly — not effusively, just genuinely. Then give a tight 60-second summary of who you are and what you're trying to figure out. This gives them context and signals that you're organized and respectful of their time.
"Thanks so much for making time for this. Just to give you a quick frame: I spent three years in FP&A and I'm now actively transitioning into analytics engineering. My SQL and business context are strong but I'm working on building more engineering credibility. I'm particularly interested today in how you think about the analytics engineering / data engineering boundary — that's a question I keep running into as I look at roles."
Then hand it to them: "Does that give you enough context? And I'm happy to take us wherever is most useful."
That last line is important — it signals that you're not running a rigid script. You're having a real conversation.
The biggest mistake people make in informational interviews is asking all their prepared questions in sequence, regardless of what the other person is saying. This feels like an interrogation, not a conversation. When someone says something interesting, follow it. Ask a follow-up. Sit with an idea for a moment. This is how you get the real insights that you can't get from a LinkedIn profile — the things people actually think, not the things that are safe to say in a structured Q&A.
If someone says, "Honestly, the biggest challenge isn't the technical stuff — it's the organizational politics around data ownership," don't nod and move to your next prepared question. Ask: "Can you say more about that? What does that actually look like day to day?"
That's where the gold is.
Take notes visibly. If it's a video call, tell them at the start: "I'm going to take notes as we go — I hope that's okay." Most people find this complimentary. It signals that what they're saying matters to you.
Watch the clock. At the 20-25 minute mark of a 30-minute call, if they haven't already started wrapping up, you should signal the wind-down:
"I want to be respectful of your time — I know we said 30 minutes. Would you like to wrap up, or do you have a few more minutes?"
If they say they have more time, great. If not, you have one final question to ask: "Is there anyone else you'd suggest I speak with?" This is the single most important question in an informational interview, and you should almost always ask it.
A referral from this person to another person is worth ten cold messages. It changes the temperature of every subsequent outreach you do. And it's easy for the person to give — they're not hiring you, just making an introduction.
Close with clarity and gratitude. No fawning, no lingering. Something like:
"This has been genuinely useful — the way you described the relationship between dbt metrics and BI tooling gave me a completely different way to think about that problem. I really appreciate your time. I'll follow up by email to say thanks properly."
That last sentence sets an expectation that you will do something after the call — which we'll cover in the next section.
This is where 90% of people drop the ball. The informational interview goes well. Both parties feel good about it. And then... nothing. The conversation sits in LinkedIn history, disconnected from anything else, and fades.
Here's what you do instead.
Send a real thank-you, not a template. Specifically reference 2-3 things from the conversation that were useful — this proves you were actually listening and gives the person a reason to remember what you talked about.
Subject: Thank you — that conversation about the metrics layer was genuinely helpful
Hi Priya,
Thank you again for taking time yesterday — that was one of the most practically useful conversations I've had in this whole process.
Two things that will change how I work: your point about treating the semantic layer as a contractual boundary between engineering and analytics is a framing I hadn't articulated before, and it immediately changed how I'm thinking about the project I'm building. And the specific advice about making sure I can speak to data modeling decisions — not just implementation — in interviews was exactly what I needed to hear.
I'll keep you posted on how the job search goes. If I come across anything in the analytics engineering world that seems like something you'd find interesting, I'll send it along.
Thanks again — [Your name]
Notice what this email does not do:
The relationship has to be maintained, and that requires occasional contact. The rule is: only reach out when you have something genuinely relevant to share with them, not just when you need something for yourself.
What qualifies as a genuine reason to reach out:
The cadence for these touch-points is roughly every 4-8 weeks for the first few months, then less frequently. You're not managing a CRM. You're having a natural human relationship with someone in your field. Let it breathe.
At some point, the relationship may reach a point where it makes sense to ask for a more direct favor — a referral, a connection to a hiring manager, a review of your portfolio. Here's how to approach that moment:
Wait until you've had at least two meaningful contacts — the informational interview plus at least one substantive follow-up interaction.
Be explicit about what you're asking and why you're asking them. Don't hint. People can tell when they're being worked up to something, and it feels manipulative. Just say it:
"Hey Priya — I have a direct ask, and I completely understand if it's not something you're comfortable with. There's a Senior Analytics Engineer role open at [Company] and I noticed you worked with their Head of Data two years ago. Would you be willing to put in a word, or give me a sense of whether that's even the right fit based on what you know of the team? No pressure at all."
That's specific, honest, and explicitly releases them from obligation. People respond much better to that than to a slow build toward an ask they could see coming.
This exercise has four phases. Complete them in order over two weeks.
Identify five data professionals you could reach out to using the criteria we discussed. For each person:
Write three full outreach messages from scratch. Before you send each one, audit it against this checklist:
Send the messages on different days, not all at once.
When you get a yes (and you will), prepare using the research process described above. After the call:
After one week with all three messages sent, review your response rate. If you're at zero for three, something is structurally wrong with your outreach. Identify the most likely culprit: Was the hook too generic? Was the ask too vague? Was your "who I am" section too long? Rewrite one message from scratch with a different approach and send two more.
This is almost always imposter syndrome talking, not reality. You don't need to be an expert to have an informational interview — you need to be curious and have done your homework on the person. The asymmetry in knowledge level is the whole point. If you knew everything they know, you wouldn't need the conversation.
That said, if you genuinely don't know enough to ask specific questions, that's a preparation problem, not a conversation problem. Fix it before you send the message: spend two hours reading about the specific area you want to discuss.
Low response rates (under 20%) after a good-faith follow-up indicate one of these issues:
The hook is generic. Re-read your first sentence. If someone else could have written it, rewrite it.
The ask is too big or too vague. "Pick your brain" or "any time you have" are asking for open-ended commitment. "20 minutes specifically about X" is a bounded, concrete request.
You're targeting the wrong seniority level. Very senior people (VP and above) have much lower response rates to cold outreach. Start with individual contributors and senior engineers — they're more accessible and often give better tactical advice anyway.
Your profile isn't credible. If your LinkedIn profile is sparse or looks like you just created it, that raises red flags. People need to be able to see who you are before they agree to talk to you. Build the profile before you build the outreach list.
Some conversations do stall. You ask a question, they give a short answer, and there's an awkward silence. Have a recovery ready: "Let me try that from a different angle — is there a specific project or problem that comes to mind where that dynamic played out?" Specificity is the antidote to stalled conversations. Move from the abstract to the concrete.
Have a 60-second pitch ready. Not an elevator pitch — a clear, honest answer to "so what are you looking for?" Practice it before every call:
"I'm targeting analytics engineering roles at companies with mature data stacks — ideally where there's real dbt infrastructure in place and the work involves more modeling and design than just query writing. I'm open to mid-size companies and Series B or later startups. I'm not ready to move yet — I'm still in conversations and building my portfolio — but I'm trying to be smart about where I focus my energy."
That's honest, specific, and tells them what you need without asking them for it.
This is a good problem to have, but it does happen. If someone spontaneously offers to put your name forward for a role and you're not ready (portfolio incomplete, interviewing skills rusty, etc.), be honest:
"I really appreciate that — that means a lot. Would you be okay if I followed up in 4-6 weeks? I want to make sure I put my best foot forward rather than having you put your name on the line for something I'm not fully ready for."
People respect this kind of self-awareness. It also means the referral happens when you're actually ready to convert it.
You're not, as long as you're doing this correctly. Most data professionals remember being where you are. Most of them got help from someone. Many of them actively want to give that help forward. What bothers people isn't being asked — it's being asked poorly. You're learning to ask well.
Let's consolidate what you've built.
Informational interviews work because they operate on human connection rather than automated application pipelines. When you reach out with specificity, conduct conversations with genuine curiosity, and follow up with real relationship-building, you create something job boards can't give you: people who know who you are before a position opens.
The mechanics are clear:
The deeper principle underneath all of this is that informational interviews teach you to see your job search as relationship-building over time rather than transaction-seeking. That shift in mindset changes how you show up — and people can feel the difference.
Build your list today. Identify five people using the criteria in this lesson. Don't start messaging yet — just build the list and find the hooks.
Read the work of the people on your list. Spend time with their writing, their talks, their public conversations. This is not optional prep — it's the foundation of everything that comes after.
Write your first message. Use the template structure, but make every sentence specific to the person you're writing to. Audit it against the checklist before sending.
Track everything in a simple system. A spreadsheet with columns for name, platform, date sent, date followed up, and outcome is enough. You don't need a CRM. You need to not forget who you've talked to and what you agreed to do.
Come back to this lesson after your first three conversations — you'll read it differently after you've felt what works and what doesn't in the real-time pressure of an actual call.
The data world is smaller than it looks from the outside. The same names come up repeatedly. People who know your name remember it. Start the conversations now — not when your portfolio is perfect, not when you feel fully ready. The readiness comes from the conversations, not before them.
Learning Path: Landing Your First Data Role