I have written to around seventeen groups about a PhD position. Plenty of those emails got no answer at all, and that is harder to sit with than a rejection.
The routine is the same every time. Read the group's recent papers properly, not just the abstracts. Work out what you could do there that you cannot do here. Write the email: who you are, what you have built, why this group, what you would want to work on. Attach the CV. Send it. Then wait. The reading is the slow part. The writing is quick once there is something specific to say.
The email itself asks for very little. A paragraph on what I have done, a paragraph on what I would want to do with them, a CV, and an offer to send the research statement and the transcript if anyone wants to see them. I am applying from Abu Dhabi to groups where nobody has met me, so that email is the whole introduction.
A week later you are still rewriting it in your head. Was the first paragraph too long? Did I open with the wrong paper? Are they traveling, is the funding gone, or was this a quiet no? Nobody has closed the door, so you keep it open, and an open door takes up a share of your attention that a closed one does not.
That is the practical cost. A yes and a no are both things you can plan around. A maybe that is really a no just sits at the top of the list and keeps you from writing to the group that actually has room.
A rejection ends all of it in two lines. One professor wrote back to say his lab was not the right fit this year, and that I should write to a colleague of his instead. He named the colleague. That took him five minutes and it made my day.
Nobody owes me a reply. Anyone running a group gets a lot of these emails, and from the other side of the inbox mine may look like all the rest of them. I know the arithmetic is not in my favor.
Still, if you are on that side: send the no. Two lines is enough. If you can point the person at someone else worth writing to, that is worth more than anything else you could put in the message, and it costs you nothing but the name.
I am still looking. Machine learning for drug discovery, starting late 2026 or 2027. My email is on the contact page, and I answer all of them.