How to get great thesis supervision

It’s that time of year again — students who will write their thesis in the coming semester have started looking for a supervisor. It’s a stressful period for most, which leads to messages like this one:

A frustrated student’s Facebook post

As a PhD student, I can’t help being annoyed to be put into a box. However, the reason for such a comment is that this person (who is just a random example for something that I see happen a lot) probably doesn’t know any better and is simply frustrated by the lack of progress with organising their thesis — which is indispensable for finishing their studies, so they’re quite right not to take it lightly. So if you’re reading the above message and think “they have a point, I feel the same way”, this post is for you. To help, I have collected common pitfalls students make in contacting me as a potential supervisor. I want to explain what they are doing wrong, and how to improve. Maybe I can shed light on some misunderstandings concerning PhD students and thesis supervision.

Disclaimer: this post is highly subjective. It contains (slightly adapted and anonymised) examples of emails I received over the past two years. Keep in mind that what you read here is only my opinion — other supervisors might not agree. Moreover, the post mirrors the way things work at my chair at the Department of Informatics at the Technical University of Munich. If you’re writing your thesis somewhere else, things might be different. Treat advice with caution and don’t follow it blindly (which should go without saying, really, but it can’t hurt to repeat it from time to time). I’d also love to hear your thoughts or suggestions to this article via email.

Your thesis from the perspective of the supervisor

An upcoming thesis tends to evoke a wide range of feelings with the author: maybe you’re excited to work on a particular topic, or annoyed that you have to write a thesis in the first place (“When am I ever going to need that skill in my later job?”). Some can’t wait to get it over with and finally finish their studies, others don’t believe they can do it and become a shaking bundle of nerves even before having started. Whatever feelings you might have, put them aside for a minute and consider the topic from your (potential) supervisor’s perspective. Let’s talk about their motivation to supervise you. You might be thinking something along these lines:

“As a student, I was convinced that it was my supervisors job to supervise me, since that was what he got paid for. I felt that I would be doing him a favor if I came up with my own thesis topic, since this would save him the effort of having to think of one. I also felt that since I was essentially putting in six months of full-time work for free, I deserved a fair share of his time in return. Not providing that supervision could then only be attributed to an unfair character of the supervisor.”

This is from Elmar Juergens’ excellent page, thesisguide.org. He goes on to explain why this view is wrong, and what the typical workday looks like for a PhD student. I encourage you to read the page on how to get great supervision as well as the rest of the content, but as a summary, let me illustrate with some humour the situation that we often find ourselves in:

Funny or depressing?

In short, a PhD student’s typical day is full of distractions from what should be your main goal: doing research to fill a PhD thesis, and writing said thesis. Yes, teaching courses and supervising students are part of your daily duties (and in my case, although this is rare, teaching is the main reason why I chose this career), but you’d ideally want these activities to somehow support the above-mentioned goal.

As a student, you have to decide what kind of supervision you want. Do you want to get your thesis over with as fast as possible, or do you actually want to learn something? Writing a thesis will develop your analytical capabilities and develop your skills in critical thinking, reasoning, and writing. A good supervisor can identify your weaknesses and help you improve in each of these. You’ll most likely not get a chance for such a close supervision in your industry career anytime soon, so you might as well take it now. Also, keep in mind that the professor is the one who will be grading your thesis in the end, based on the supervisor’s suggestion as well as their own impression of your thesis and presentation. Your supervisor knows what the professor likes (and hates!) in a thesis, and can give you valuable advice that will make your thesis better. Of course, you can much more easily get a supervisor that you meet once for registering the thesis, and once for handing it in, but I guarantee you that the experience will be less than pleasant and you’ll be none the wiser afterward.

As you might imagine, this kind of “good” supervision as takes time and energy. If this is what you are aiming for, you need to understand and respect your supervisor’s perspective in this matter to make sure you do whatever you can to motivate them for the job. Making your topic relevant to their research should be your top priority, but their motivation is not the only reason for doing so. If you pick a topic that your supervisor is already well-acquainted with, you will also benefit from having a better defined scope of the thesis, because they will be able to set clear goals and give you a starting point or a basis (in the form of existing research or tools) on which you can build.

If I supervise a thesis, I like to meet with the student every week, or at least reserve a fixed time for them in case they have questions or want to show me the current state of their work, even if it’s just to clear up uncertainties or give a bit of encouragement at the right moment. In addition, I read one or two of their first chapters and give in-depth comments on their style of writing as well as their flow of arguments, citation style, and much more, so that they can learn from the feedback and improve not only those passages but also the rest of their document (see also this article on an author’s and supervisor’s responsibilities). This means that it’s not reasonable for me to supervise more than around 3 theses in parallel, and I need to pick them carefully because I get much more requests for supervision than that. I only give out topics that are truly relevant to my research, and also make sure that each thesis I supervise is in the hands of a student I know to be methodical enough so that I can trust their findings, and possibly even publicise them after their thesis. This might be different if your supervisor is a Postdoc or a Professor, but in the case of PhD students, this is the only approach I know for supervisors who want a close relationship with their students, simply because there is no other way to tackle the daily work you have to do on top of your research that you want to do.

The good, the bad and the ugly

If you’ve read through the previous passages, you should have a clearer impression of what makes PhD students want to supervise a thesis. Now armed with that knowledge, how do you use it to formulate an email with an offer they simply can’t refuse? Let’s look at how not to do it first. I scrolled through the emails asking for supervision I received since I started and clustered them into types so I can point out their weaknesses.

I’d like this to be very clear: This is not about criticising or even mocking anyone. I just find that sometimes a well-picked anti-example helps more than a simple checklist of what to do. It will hopefully also help understand the issue from a supervisor’s perspective and reduce the number of frustrated messages about PhD students who “don’t even bother replying”, like the one at the beginning of this article. The texts are slightly adapted and of course completely anonymous.

The General Inquiry

Dear Ms. Dzvonyar,
Are there any theses that your Chair is offering for the next semester? I did not find any topics that interest me on the website.
Yours sincerely,
 XYZ

While I’m sure this question is generally not sent with a bad intention, I still haven’t found out what answer people expect to this. Let’s say that I get on average 1–2 requests for supervision per day. If I were to answer them with more than a simple ‘Yes’ or ‘No’, I would have to compile and send a description of the research fields of not only myself but also all of my colleagues, along with possible topics for theses. I would pretty much not do anything else all day. You did not even send me your interests, so how am I ever going to be able to satisfy your need for information?

Also, consider this: if there were any concrete, well-defined topics for theses around, they would already be on the website. In fact, many supervisors don’t advertise topics because it takes a lot of time to put together a comprehensive summary. And if you didn’t even take 15 minutes to find out what my research is about and how it would interest you, why should I spend an hour describing each possible thesis topic in depth?

The Referral

Hi Dora,
My friend Peter wrote his thesis with you and enjoyed it, so I thought I’d ask if you have any open topics. Would love to chat!
Cheers,
 XYZ

I’m glad Peter enjoyed writing his thesis with me, and it’s great that the two of you are friends, but that doesn’t mean we are. It’s actually a smart idea to talk to fellow students to find supervisors, but that doesn’t spare you the effort to at least briefly familiarise yourself with that person’s research and find out what possible topics would fit in there and whether you would even like to work in that field. If you simply put a sentence about your friend’s thesis in there, along with 1–2 aspects that you found fascinating or that were left open for future research, you’re already off to a much stronger start and you haven’t even invested more than a couple of minutes of your time.

The 20-Second-Email

Hi Dora,
I took your lecture last semester, and liked the way you teach. I want to write my thesis soon. Please inform me about possible topics and if I find one that awakens my interest, we can arrange for a meeting to discuss further.
Cheers
 XYZ

This one is really a mixture of the General Inquiry and the Referral, phrased with a real sense of entitlement. I get that you want to work on a topic that you find interesting, and I appreciate the flattery (after all, teaching was the main reason for me to start a PhD so I like to hear that I’m good at it), but ask yourself this: if this were a fellow student asking you to collaborate on a study project, how would you react? I like to have interactions with students on the basis of mutual respect, while this email basically says “amaze me, and I’ll see if I will work with you”. You should not communicate to people this way, regardless of their status. You’re not a venture capitalist — I don’t need to “pitch” thesis topics to you so you can decide which one you “invest in”.

The Industry Buzzword

Dear Ms Dzvonyar,
I am writing to you to inquire whether you would supervise my thesis. I am currently working in the company ABC, and they are interested in supervising a thesis from the field of [insert generic buzzword here, presumably containing the terms ‘IoT’ or ‘Lean’]. However, they leave me complete freedom to choose the specific topic. I would like to meet with you to discuss possibilities.
Thanks in advance and best regards,
 XYZ

Now we’re getting more into matters concerning content — or, in this case, a lack thereof. The general problem with external theses is that there is one more stakeholder than needed. This is a good moment to think about motivation again: a company most probably wants you to write their thesis there for one of the following reasons:

they want a job done for little money (in computer science, it’s often a piece of software to be developed), they want to try out something (e.g. a new technology) without risking too much; in the case of failure, they did not sacrifice the work hours of a full-time employee who has much higher opportunity costs they want to “test you” to see whether they offer you a job after your studies, or if you’re already working there, they want to keep you in their company so they can employ you full-time after your thesis.

While these are all valid reasons, they are fundamentally in conflict with what the university, i.e., your supervisor and the professor, expect from a typical thesis. In most cases, this leads to problems and ultimately to either more work for the student, or a worse grade, or both. Even though the company says that they are completely open on the topic, at some point this will change. I have seen it happen multiple times, and the student always ended up doing twice the work to basically produce two completely different representations of the same project. I can only advise you to think twice about writing an external thesis and if you do, to make sure all stakeholders agree on the scope and requirements upfront. As a supervisor, I would not accept such a thesis unless I can set the expectations so that they fit very well into my research. In certain situations, it can very well be beneficial to supervise such a project, for instance to evaluate an existing, academically proven system in an industry setting.

The Shot in the Dark

Dear Ms Dzvonyar,
My name is XYZ and I Study Computer Science in the 5th semester Bachelor. I want to write my thesis in company ABC in the upcoming term and am looking for an advisor. In the following abstract, I shortly describe what it is about and why I think it relates to your research fields.
[insert generic, buzzword-laden abstract here with no justification of how it builds on existing research, and a lot of irrelevant, but supposedly attractive details about the company at hand]
I am looking forward to your response. Best Regards,
XYZ

See above for the problem with external theses. The bigger problem here, however, is that the author of this message has obviously sent out a mass email without doing their research first. A simple look at the publications of a potential supervisor would have made it clear that this topic is not a match. Also, unless you work in their HR department or get a 4-digit bonus for referring people, it’s not worth trying to “sell” the company to me.

The Super-Defined Project

Dear Dora,
I am currently in my final year of Masters and am looking to start working on my thesis shortly. I have been working at startup ABC for the past two years and they have proposed the topic “Development of a [insert workflow that a typical tech company uses every day] Tool for [insert industry name]”.
The topic largely overlaps with what your Chair is doing. I will be working with [insert Framework name here] and I will develop in [insert programming language here]. Most probably, I will compare [insert list of already existing tools or frameworks] in terms of their functionality and fit to the problem at hand.
I would like to arrange a meeting with you to discuss the topic further. Best Regards
XYZ

Now we are talking about the other end of the spectrum compared to the Industry Buzzword, and this email illustrates again the fundamental issue with external theses I described there. I can already see that this thesis is a ticking time bomb of conflicting expectations and double work. If your company is not even open in terms of technologies and frameworks involved, how will you as an author ever have the chance to methodically examine the topic, an activity that usually requires developing and investigating hypotheses, considering alternatives, and evaluating possible solutions? A thesis has to have academic value; otherwise, you’re nothing but a coding monkey, and an underpaid one for that matter. Sure, the topic overlaps with what we are doing (we are called the Chair for Applied Software Engineering after all), but even if you are making a good case for why said tool is needed, I would not supervise a rigidly defined project like that, as I’m sure most colleagues would agree. My recommendation is to tell your boss to hire you part-time to develop that system, and to go write a thesis directly with a university supervisor while you’re working there. I did that during both my theses and with a bit of sensible time management, it worked out just fine.

The Transcript (added 03/2018)

This one does not come with an email text, just a general piece of advice: you can choose to send your prospective supervisor a transcript, but if you do, make sure that your grades in the field at hand (and especially the courses given by that professor) are good! I’ve had several students send me their transcript without me even asking for it, and their grades in all Software Engineering lectures were well under average. Everyone has different strengths and interests, and that’s okay — but what do you want to express by voluntarily sending me a transcript that essentially shows that you are generally able to obtain good grades, except for the field you claim to be exceptionally motivated to write your thesis in? Obviously, there is a wide variety of reasons other than a lack of interest that can explain your grades, but if you think I should see your transcript and it’s not self-explanatory, address this obvious fact.

Get to the point, please.

Yes, I’ve been rambling. I have deliberately shown you anti-examples in this post, to try to convey how a PhD student might perceive a certain email you send them. All this is written with the best intentions: to help you improve in targeting potential supervisors and to ultimately end up with a thesis and supervisor you are happy with. This is what I would recommend, in a nutshell:

  • If you know someone who is a PhD student, Postdoc, or even Professor from a previous course, that is always a plus. Go to them personally. I would try to arrange a meeting via e-mail first, and if that does not work, try to knock at their office door, politely ask them for a conversation, but also be respectful if they are currently busy and ask when would be a better time. Of course, it helps if you performed well in their course, stood out by asking questions or solving exercises, or if you attended a “cosier” course format like a small seminar from which they still remember your name.
  • Regardless of who you are contacting, do your research. You don’t need to cyber-stalk each person you write to, but it’s only reasonable to expect that you have invested 15 minutes of time familiarising yourself with your potential supervisor’s interests. Here are some tips on where to start:
    1. Check out their profile on their chair’s website (many people list research interests there).
    2. Take a look at their publications: if they have not listed them on their website, it’s quickest to use the Google Scholar Profile Search or simply enter their name on Google Scholar.
    3. If they’re a Postdoc, read the introduction to their dissertation (since those are always published, you can download it from the university’s library or search for it online).
    4. Skim through the most recent theses they supervised; concentrate on introductions and future work.
  • Build on what you know about their work and share a couple of concrete ideas on what you would be interested in and why, but leave the final topic open for discussion. Describe your prior experience in that field, if any.
  • Tell them the timeline of your thesis: when do you want to start? Are there any constraints, e.g. you can only start after a certain date, or you have to be done by a particular date?
  • Attach relevant information to the email you are sending. I usually ask for a piece of academic writing. Whereas to me, the student’s grades don’t matter so much, other supervisors might like to see an academic transcript or a CV. Either way, whatever you send should support your message, don’t attach random documents to your email because you think they might interest someone.
  • This should go without saying: grammatical errors and typos are a no-go. Basic proofreading is really not too much to ask.
  • While this section is admittedly very long, your email shouldn’t be: keep it simple, to the point, and only share information that is relevant.

This is all the advice I have. Sure, even if you follow all of this, it might still be that a supervisor still doesn’t accept you simply because they have too many theses, an important project that is keeping them busy, or something else that is standing in the way. But if that happens, you’re no worse off than before.

I invite you to try this and tell me about the results. Also, feel free to share any feedback you have on this post through any of the channels available at the top of this page, whether it’s praise, criticism, a rant, suggestions for more anti-examples or new content of any kind. I would also love to hear whether these principles apply to other universities and different fields. I will add feedback I get to this post, so it will evolve over time. For now, good luck with finding a supervisor and “have fun” writing your thesis! ;-)