Most early researchers think collaborations are about asking for help.

They’re not.

They’re about creating value—together.

And how you approach a potential collaborator will either open doors… or quietly close them.

I learned this the hard way…

Back in the early days, I’d write emails that went something like this:

“Hi, I admire your work. I’d love to collaborate sometime. Do you have any projects I can help with?”

Most of the time? Silence.

Occasionally, a polite—but noncommittal—response.

I didn’t realize it then, but I sounded vague. Passive. Like I needed a hand, rather than someone worth bringing onto a project.

Everything changed when I flipped the script.

Here are the 6 collaboration rules I wish someone had handed me when I was starting out 👇

1️⃣ Lead with value

(Everybody’s favorite radio station is W.I.F.M—What’s In It For Me.)

Before you reach out to anyone, pause and ask:

“What exactly am I offering them?”

Because no matter how brilliant someone is, they’re not scanning their inbox hoping for another request. They’re looking for momentum. And clarity.

That means no more “Let me know how I can help.”

Instead, offer something concrete. A cleaned dataset. A first draft of a figure. An R pipeline they can easily plug in.

Here’s one approach that worked well for me:

“While analyzing [X dataset], I found a pattern that reminded me of your 2021 JAMA paper on [topic Y]. I built a draft figure exploring this angle. If you’re interested, I’d love to co-develop it further.”

You’re not just saying “I admire your work.” You’re handing them a springboard.

That gets replies.

2️⃣ Be the kind of collaborator you want to attract

This one sounds obvious—but it’s not.

Ask yourself:

• Would I want to work with me on a deadline?
• Do I follow through, or need to be nudged three times?
• Do I send polished drafts—or ask others to figure things out for me?

Before I started getting yeses, I had to level up quietly.

I took courses in R. I practiced writing concise, journal-ready paragraphs. I learned how to clean messy EHR data without pinging a busy collaborator for every decision.

That’s the difference between “interested” and “ready.”

So if you’re reaching out to someone you admire, make sure your skills and mindset match the energy you’re asking them to invest.

Because the best collaborators aren’t looking for more students to supervise. They’re looking for peers they can build with.

3️⃣ Start with your circle, then go wide

Don’t make your first ask a cold email to an Ivy League PI.

Start closer to home.

• A fellow resident who’s doing research in your area.
• A senior at your institution with a skillset you don’t have.
• Someone you met at last month’s journal club or webinar.

One of my most impactful collaborations started with a short chat at a conference poster session. We realized our datasets had synergy. I followed up with a shared Google Doc and a draft abstract. That one conversation eventually led to a first-author publication.

Here’s the truth:

Warm intros and shared spaces beat cold emails every time.

But if you do go cold, make it specific and useful (see rule #1).

4️⃣ Collaborate across fields

I used to think “collaboration” meant “find another person doing what I’m doing.”

But some of my best collaborations came from people doing something entirely different.

Like the bioinformatician who helped me run PCA on a clinical dataset. Or the machine learning expert who taught me how to balance model performance and interpretability.

These weren’t one-off favors. They reshaped how I think about research questions.

Because people outside your field won’t just help you answer questions.

They’ll help you ask better ones.

So if your work is starting to feel narrow or stale, look laterally.

Your next great idea might come from someone who reads a completely different set of journals.

5️⃣ Be project-specific, not vague

“Would love to collaborate someday” sounds polite. Harmless.

But it’s the collaboration-killer.

Because it gives the other person nothing to latch onto. No timeline. No topic. No direction.

Compare that to:

“I’m working on a model of treatment response in PsA using [data source]. I saw your work on phenotype clusters and wondered if we could explore a joint angle.”

Now you’ve done three things:

1. Shown you’ve done your homework.
2. Connected your current project to theirs.
3. Made it easy to say yes—or suggest a next step.

Specificity doesn’t limit your options. It creates momentum.

6️⃣ Build slowly, then go deep

Don’t chase 10 collaborators at once.

Chase depth.

Here’s what that looks like:

• You start by sharing a figure.
• Then maybe you draft a paragraph.
• Then you run some joint analysis.
• Then it turns into a full manuscript.

At some point, you’re not just co-authors. You’re thought partners. You know each other’s strengths, blind spots, and working rhythms.

That’s where the real magic happens.

Because over time, good collaborators become the people you can text when a reviewer trashes your methods—or when you need a sanity check at midnight.

That kind of trust doesn’t scale fast.

But it’s worth investing in.

This week’s action step:

✅ Make a list of 3 researchers whose work you respect.

✅ Find one concrete intersection with what you’re already working on.

✅ Draft a short message that leads with value—not vague interest.

Then hit send.

And remember:

Collaboration isn’t a transaction. It’s a relationship.

Start it like one.

Which non-obvious field could bring a surprising new dimension to your research – who’s the first person you’ll reach out to today?