The Academic Job Search Epic

A few months ago my (first) academic job search epic came to an end. In January 2016 I will start the Kryazhimskiy Lab at UCSD, in the Section of Ecology, Behavior and Evolution. I thought it might be useful for the next generation of applicants to summarize my experience. So, here we go.

Some stats

Let’s start with data.

2013 2014
Applications 20 29
Interviews 1 11 (+1)
Offers 0 3 (+1)

I applied for jobs two seasons in a row. I started applying in the fall of 2013 and continued throughout the fall of 2014 (I also sent out a very small number of applications in 2011, but I was clearly not ready then). My applications in the two seasons were essentially identical: same CV, same recommendation letters, same ideas in the research statement. For complete disclosure, I did rewrite my research statement in 2014 and I do think it became better. The main difference though between applications was a single line in my publication list. In June 2014 our Science paper came out. Importantly, by the time of my 2013 application, all experiments were finished and data were largely analyzed and described in my research statement. In other words, my case is a nice test of the effect of a single Science paper on academic job search performance.

The effect is striking. In 2013 I sent out 20 applications and got 1 interview. It is at that time that I wrote this depressing post. In 2014 I sent out 29 applications and got 12 interviews, one of which I declined. In fact, there were 4 departments to which I applied both in 2013 and 2014, and I got an interview at 2 of them in 2014 but none in 2013. As a result, I got 3 offers in 2014. One additional university was prepared to give an offer, but came through too late in the game. From what I hear from other colleagues who recently applied for jobs in biology, an overall 10% success rate is pretty reasonable.

What do we learn from here? I think two lessons.

First, it appears to be very important to have a major paper from your postdoc published. Not in preparation, not on bioRxiv, not in review. Published. Preferably in Cell, Science, or Nature. I guess we all know that by now.

Second, even if you have a major paper out, you should send 10+ applications to ensure a decent chance of success.

There are certainly exceptions to these “rules”. I think there are essentially two “tracks” for getting a job. On the slow track (which is what I did) you do a regular postdoc (or multiple), where you obtain a solid if not stellar publication record and get your name recognized in the field. To start on the fast track, you have to have a stellar PhD (i.e., one or more high-profile papers). To continue, you need to get a prestigious independent postdoc position, e.g., Harvard society fellowship, Miller fellowship, etc. Then you might be able to get a tenure-track job after just one or two years of this super-postdoc even without major publications.

Time investment

One important thing to consider in applying for jobs is time investment. Putting together an application takes a lot of time. I spent 2 to 3 weeks working full time on my research statement in 2013 and another 2 weeks in 2014. Plus one week on teaching statement. Plus 3-4 days crafting an effective cover letter. You may be able to do better than that if you are a fast writer, but even then conceptualizing your research program, coming up with specific projects, and going through two or three feedback/re-writing iterations still takes time. Adjusting your statement to different page restrictions is another time-sink factor. Finally, you might want to tailor your statement to different types of departments. For example, I applied to ecology and evolution departments and to systems biology departments. Even though all core ideas and proposed work of course remained the same, I did frame them in different ways.

Once I had a couple of versions of the research and teaching statements and the cover letter backbone, it took me about 1 day per application. I looked up the department and carefully re-read the job ad trying to understand what kind of person they might want. So, I tried to place appropriate accents in the cover letter and to a lesser extent in the research statement to make my application slightly more appealing to a particular department. I have no idea whether it played any role or not. According to this post, it might. If I read my application, I would probably conclude that the person took the application process seriously.

Given this considerable time investment, the major question is: is it worth it? If you have a good chance of getting a job the answer is obviously “yes”. The answer is not necessarily “no” even if think that your chances are not great. There are three reasons for that. First, you never know your chances until you try. Second, if you get even just one interview, it is a very valuable practice for the future. Third, even if you do not get any interviews, going through the process of writing the research statement is not a complete waste of time. Conceptualizing your research program takes a lot of time, and I truly believe it is helpful to start writing early. My current research statement is compelling, but the one I wrote back in 2011 is much less so. And I arrived at my current level of conceptualization through multiple iterations which were spaced out by several months and interspersed with reading and digesting new papers and talking with smart people. In short, writing your research statement is a fairly useful time investment. However, it has to be weighed against the time investment into actual research and paper production. And the equation is simple. If you have awesome papers, you will likely get by with a poor research statement (I know of such precedents). If you don’t have papers, a fantastic research statement will never get you even an interview.

A couple of notes on how I wrote my research statement

There are probably good guides for writing research statements out there, but here are my five distilled ingredients.

  1. Vision. In my mind, vision encompasses a deep scientific problem and a series of approaches to attack it. This is the glue that holds the whole statement together.
  2. Focus. I tried to structure my statement so that it feels like every proposed project fits with the vision rather than goes off in a different direction. The reader should feel that if you accomplish all these projects, the field will advance in a major way.
  3. Clarity. Keep in mind that the first reading will unlikely take more than 10–15 minutes. So, better to assume too little knowledge on part of the reader than too much.
  4. Expertise. I tried to demonstrate that I have enough expertise to accomplish each proposed project. In parts where my expertise was clearly lacking, I explicitly mentioned relevant collaborators at the particular institution.
  5. Scope. I was given the following advice. There have to be enough ideas in the proposal to fuel a lab for five-seven years. Estimate how much work each project will take. There should be more work than you can accomplish yourself, but not so much that it will require a whole institute.

On the importance of networking

Everybody says that networking is very important, but it is actually pretty hard to pinpoint how it actually plays in. Here is my data.

Departments Got interview Did not get
With direct connection 7 7
Without direct connection 5 10

I am showing here the numbers of departments to which I applied in 2014 stratified by the presence of at least one faculty member who knew me personally (“with direct connection”; NB: I have no information on whether this person was on the committee or not) versus no such faculty members (“without direct connections”). So, it looks like the chances of getting an interview at a department with a direct social connection are a bit better (50% vs 33%). Still, publishing (at least in high-profile journals) is a much better way to improve chances of success than networking.


My interviews began in November and finished in March. There are a couple of interesting things that I learned, both about myself and about the process.

First, the delivery of my talk varied a lot, and the only predictor that I could come up with is whether there were people in the audience who intimidated me on a purely subconscious level. The presence of such people sometimes caused me to fluster. I noticed that this behavior was aggravated by lack of sleep suppressed with coffee. The lesson is: Your body and mind can behave in unusual (annoying) ways under stress, lack of sleep, and various kinds of intoxication. These undesired behaviors can be controlled, but it is important to pay attention and not to aggravate them.

Second, the variation in the quality of the departments is staggering. I applied only to places where I could potentially see myself work and my family live. Even so, the departments where I interviewed varied from absolutely stellar (which is what I got used to during my PhD and postdoc years) to I-would-rather-quit-science-than-go-there kind of places. It’s good to keep in mind that before a physical visit you have no idea of what the department is like. Some places that you think are not even worthwhile applying to may turn out to be really great, and vice versa.

Third, my ability to predict after the interview whether I would get an offer was close to zero. The only exception were failed interviews, i.e., if I thought the interview went bad, it most certainly did. The converse was not true at all: interviews that I thought I nailed did not necessarily result in offers. After each interview, just for fun I wrote down my perceived probability of an offer (either 10%, 25%, 50%, 75%, or 90%). The average value of this probability for places that did indeed offer was 63% (±20%); and for places that did not offer it was 44% (±15%). So, barely better than coin tossing.

After the offer

I was lucky enough to receive the first verbal offer after my second interview in early December. You might think that this alleviated much of the stress of subsequent interviews, but it actually did not because the offer was from the place that was not number 1 in my internal preference list (which was misguided, see above). Moreover, when most of the interviews started, the soothing effect of the offer wore off. By January it seemed more like a distant dream than reality, and since it was a verbal offer I had no document to convince me otherwise.

In any case, in the first week of March I received two more verbal offers. After that the negotiation process began. There are several important things that I learned in this process. First, until you have a written letter in your hands, nothing is certain. I have heard of multiple cases when the unofficial offer was withdrawn. So, work towards obtaining a written letter from the first phone conversation.

Second, as in any negotiation process, it’s very important to know what you want. You have to have a very clear hierarchy of priorities. And I think it is best to put out all the “deal breakers” right out on the table. This way, it will be no surprise and hopefully no bad blood if you fail to come to an agreement.

Third (and this is an advice from a senior colleague), if you have multiple offers, you have one “trump card”. You can say “If you make X happen, I will come here”. It is likely that X will happen if it is at all in Chair’s power. Obviously, you can only say this to different places with which you negotiate sequentially.

Fourth, the success of your negotiations depends heavily on your Chair. It’s Chair’s job to convince the School (the entity that has the money) to satisfy your needs and to do other things that go beyond the University (e.g., contacting potential employees for your spouse). If the Chair is doing a great job, you will receive positive enthusiastic responses to your requests. The enthusiasm of responses is probably the best predictor of how the school and the department will treat you if you actually join. If you have to beg for every single thing (lab space, startup amount, equipment, job for spouse, housing, etc), chances are that you will have to swim against the current during your whole appointment, and you won’t get tenure without competing offers. At that point you will probably ask yourself: would I really want to be in such a place?

That’s almost all I wanted to say about my own experience. Good luck to all of those who are embarking on this miserable journey this year!

NB: I was asked a couple of specific questions which I did not answer above. Here are my answers

  1. Where did you find job ads?

I used alerts on Nature Jobs and Science Careers, I was subscribed to EvolDir, I occasionally looked through postings on AcademicJobsOnline and Vitae. In my experience, most job ads are posted in multiple venues, so missing a relevant ad (at least for US-based jobs) is almost impossible. I also was incessantly looking (and contributing) to the Ecology Wiki, which I highly recommend for maintaining sanity.

  1. What would you do differently?

If I knew my chances, I would not have applied in 2013. This would have saved some of my nerves and a lot of my time.

  1. How important is teaching experience?

Not at all, at least at places where I applied (R1). But you should still put some time in writing a good teaching statement.

  1. Is outreach important?

No, unless they specifically ask for it.

UPD (27 Aug 2015)

5. What did you do after the interview? Did you follow up or just wait for an offer?

All places where I got offers contacted me themselves. When I got the offers I informed the search committee chairs of places that have not decided yet and that I would seriously consider if they did offer.

6. What were the red-flags that led you to identify “I-would-rather-quit-science-than-go-there kind of places”?

The following questions might help.

1) Do you sense or know of any inter-departmental conflicts? 2) Are theres some awesome people, or does the place feel “stale”? 3) Are there any people who you could collaborate with (or at least talk to about your science) or will you be completely isolated? 4) Are faculty excited about their grad students?


21 responses

  1. This is a very interesting read! Ugh, the job search is indeed a miserable journey. I can only echo the advice to wait until your paper is published–advice which is somehow commonplace and yet still commonly ignored. For instance, I ignored it. Ultimately, I am very fortunate to have ended up at a great place that I love, but the path might have been a bit more pleasant if I had a paper out, as I think your experience nicely documents.

    • Thanks a lot, Sergey! Is fund proposal important? Heard successful funding experience is as important as or more than publication records.

      • I had independent funding in 2013 (BWF Career Award at Scientific Interface). Apparently it’s not enough w/o a paper.

  2. Pingback: Landing an TT job | Mistress of the Animals

  3. Hi Sergey,

    This is an excellent post.

    I somewhat disagree with the necessity of the Nature / Science / Cell paper. Certainly papers in these journals help, but the most important thing is to have 1 to 2 first-author publications from both your PhD and postdoc. In 2013 you only had one first-author “experimental” publication from your work in the Desai lab (Evolution, 2012). And even though it’s a nice paper, if your postdoc publications are in less well-known journals like Evolution, committees usually want to see several such papers. In 2014, you had another first-author publication (2014, Science) from your postdoc. We’ll never know, but I bet if that paper had been in eLife or PLoS Biology or PLoS Genetics or Mol Biol Evol you might still have done pretty well.

    As an argument for this, there are several postdocs from Harmit Malik’s lab here at the Hutch who have gotten good faculty positions in the last few years with eLife and/or PLoS Genetics papers. Also, two recent hires at the Hutch (Rob Bradley and Trevor Bedford) had no Nature / Science / Cell publications at the time they applied.

    The other thing I would add is that you never really know what is going on with any given search. For instance, maybe they recently hired someone working in your area and there is an implicit bias towards some other subfield. It’s almost impossible to know these things a priori, so this is another reason why your suggestion for 10+ (I would say 20+) applications is a good idea.

    I agree with you that premature applications can be a waste of time, as you spend months writing applications, don’t advance your research, and still don’t get a job.

    Also, at least in my opinion, it’s a waste to spend a lot of time writing reviews unless you already have a few good first-author publications. At least when I look at an applicant CV’s, I completely ignore reviews and mostly ignore publications where the applicant isn’t at least co-first author.


    • Jesse,

      Well, of course a Nature/Science/Cell paper is not a formal requirement. But what you are saying about 1 Nature/Science/Cell paper vs multiple Evolution papers makes no sense to me, given how arbitrary the success of publishing in the big journals is.

      Also I feel like interdisciplinary applicants like myself (or Rasi — new hire at the Hutch) have to be considered a bit differently than non-interdisciplinary ones. Yes, sure, my second (Science) is better than my first (Evolution) paper along several dimensions and perhaps I agree that the Evolution paper alone is not enough. But I do have multiple good first-author theory papers from my first postdoc. I think it is stupid (not to mention unfair) to discount them, even if I am proposing to do experiments [I can elaborate on why I think so]. In other words, if the committee is looking to hire an interdisciplinary person, they should expect to see trade-offs in their publication record. Don’t you agree?

      I agree with other points you make.

    • Jesse,

      I like your alternative explanation, and I am biased in its favor; I want it to be true.

      Anybody who read carefully Sergey’s research statement in the fall of 2013 would have understood that Sergey has led two projects in his postdoc: both projects were completed, and one was published. The description of the unpublished project was good enough for people to get a sense for it and certainly more detailed than the two extra lines in his CV added in the summer of 2014. This suggests that at least the passage through peer review — if not the exact journal — have improved Sergey’s chances for getting a job. Even this more favorable interpretation suggests the search committees tend to outsource at least some of their critical thinking and evaluation to anonymous and mostly non-traceable process.

      • To answer Sergey’s and Nikolai’s posts:

        Sergey’s future research clearly has a major experimental angle. So I don’t think it is “stupid” for committees to want to see published experimental papers.

        In addition, my guess is that most of the jobs to which Sergey was applying were closer to general biology positions than theoretical evolutionary biology positions. The truth is that Evolution is not a well-known journal (compared to say PLoS Genetics or eLife) outside of the theoretical evolutionary biology community. So it would be quite easy for a general biology search committee (when going through a long list of applications) to look at his CV in 2013 and think, “this is all purely computational / theoretical work.” (And by the way, I think that purely computational / theoretical work is great; my favorite paper of Sergey’s is his 2008 PLoS Genetics dN/dS study — it would have saved a lot of hyperventilating about Ebola’s evolutionary rate over the last year if more people thought about these results. But again, I’m assuming that most places to which Sergey applied were doing general biology searches rather than theoretical evolutionary biology searches, and that his proposed research involved a lot of experiments.)

        As far as describing his unpublished work in his application, the truth is that unless someone is EXTREMELY familiar with ongoing work, it’s hard to get the benefit of the doubt from search committees (or grant-review committees) if there isn’t some accompanying already-published work. As Nikolai says, in some ways this implies outsourcing evaluation to the peer review process. But it’s impossible for a search committee to peer review all the unpublished research of hundreds of applicants in diverse fields. So the process probably works like this: (1) If it hasn’t passed through peer review, it’s mostly ignored. (2) If it has passed through peer review and the rest of the application looks good, then the committee reads the published paper to form their own opinion.

        Anyway, overall I agree with Sergey — the most important thing for getting a job is a good publication record. I just don’t think that these papers need to be in Nature. They just need to be in a journal that has reasonable name recognition among the people on your search committee. If you’re applying to a general biology search with an evolutionary slant, there are lots of journals that meet this bar (such as eLife, PLoS Genetics, PLoS Biology, Mol Biol Evol, etc).

        I’m not saying that the current system always finds the best applicants, but it’s rooted in some reasonable notions. If someone doesn’t publish papers at a pretty good rate, they’re going to have trouble succeeding as faculty. If someone is proposing experimental research but hasn’t published much in the way of experimental papers, a committee has good reason to be skeptical (or at least wait until next year when they’ve proven that part of their track record). Sure, it would be great if we could prospectively identify people with good ideas BEFORE they passed through all these hoops of doing the experiments, writing the papers, publishing the papers, etc — but that’s extraordinarily hard to do in practice.

      • I don’t want to belabor my particular case here: there certainly are objective reasons for decisions that committee members make. But even when the actions of individuals are perfectly rational, the overall group dynamics is suboptimal, as far as I can see. In any case, I think a further discussion warrants a beer.

  4. Sergey,

    Wow this is a fantastic post. I would have loved to see information like this while I was on the job market.

    ALTHOUGH… Like you, I applied to a small number of jobs in 2013 and a larger number of jobs in 2014. Like you I failed in 2013 and succeeded in 2014. But unlike you, I never had a fancy C/N/S publication. In fact, my experience is very different from yours because I achieved the unthinkable, a job without any first-author experimental papers from my postdoc! (And no, I did not have a prestigious super-postdoc fellowship, I had regular postdoc-fellowships and I had been a postdoc for over four years by the time I applied the second time.)

    For me, having connections was absolutely key. And by connections, I mean people with existing interest in collaborating with me or an experimentalist like me. I got interviews at five places over the two years I applied, and four of those were because someone on the search committee knew my work and was specifically interested in it.

    And finally, I’d like to throw out a different strategy in terms of whether/when to apply. My strategy, and the strategy I’m guilty of proselytizing to my poor friends who are now postdocs on the market, was/is to plan to go on the market for three years. First year: dip your toe in the water, apply to a couple places you’re really excited about. You never know, you might get an interview (I did!). In the following year, make sure your CV improves, and with your increased confidence and experience, hit the market hard and apply broadly. Plan to do this for two years in a row, because what jobs are out there and what they’re looking for is beyond your control. Finding ways to stay in a postdoc for those three years is MUCH more in your control–really hard for some people, absolutely, especially those with families–but for me finding ways to alleviate the incredible stress of things outside my control was very important.

  5. Pingback: Chalk talks are awesome! | Slavov Lab Blog

  6. Pingback: Chalk talks are awesome! | Forest Vista

  7. Hi Sergey,

    Thanks for the post!

    You are talking about a lot of interesting aspects on being on the job market (and the comments below make this discussion even broader). You said that knowing the future, you would not apply the first time. Most of the people I talked to had a different opinion. Wouldn’t you agree that by applying the first year you got a lot of experience and intuition about what to pay attention to during your interview? Do you think you performed differently the second year?

    The most common comment that I hear is that you simply have to be lucky. And there is not much you can do about it. As you said, you don’t know anything about the place until you go for an interview. Some people told me that you are not going to get an offer if there is another person working on a similar subject, but I know some people who got an offer because there was already a group of PIs working on a similar thing and they wanted to make a strong group in field X. I know one committee member who said that they don’t open the application if they don’t recognize the name. Many people told me that having your personal website is a must in case the committee googles your name. I also heard that it matters how many postdocs you did. In theoretical physics, it is common to have 3 postdocs. In bio sciences most of the people will apply after one longer postdoc and I heard that doing a second postdoc in a big risk (unless it is a very different postdoc). In general, my feeling is that the departments are looking for people doing interdisciplinary research, preferable theory and experiments. The problem is that they want you to reach same level as people doing only theory or only experiments. And all of the mentioned aspects matter (papers, research statement, connections, fellowships,…), they just have a different weight depending on the place you are applying. The conclusion seems to be that you should do the best that you can and hope for the best:)

    • Every case is different, and I would not generalize too much from my particular experience. I do think though that the tradeoff I describe is real, and each person has to decide for themselves what is more important for them at that point: proposal writing + interview experience versus progress in research.

      As for other factors, I agree that having a website is important these days and that having multiple postdocs is a minus (unless they are different and both show substantial results). In general, each year post-PhD is counted against you, and you have to justify each year spent as a postdoc with an eventual publication (again, with the exception of people on the fast track, it seems).

      As for interdisciplinary research, this is actually a very important misconception. There is a lot of talk about “interdisciplinary” research, or “quantitative” biology, etc., but such interdepartmental searches tend to fail more often than regular searches (I was part of multiple such searches). There is a lot of disagreement between committee members in such searches. People with different backgrounds have different views on what directions are important, they recognize different sets of names, they have different expectations for what constitutes a good publication record, etc. In other words, it seems particularly difficult to satisfy such interdisciplinary committees.

      • I totally agree with you that every case is different. Part of it is the place you are applying to (e.g., PUI vs R01). I also agree that you have to decide what is more important for you at a given time – progress in research or nicer proposal. I was just trying to ask whether you find any advantages in going through the application process twice. Most of the people I know didn’t get any offers the first time (but to be fair, maybe they didn’t try as hard as they could, knowing that they are going to apply again next year anyway).

        I agree with most of what you said about interdisciplinary research. For me, the biggest problem is that everybody talks about quantitative biology and merging model and experiments, but their definitions vary a lot. For example, I hear very often people talking about “having a model”. When I hear this, I think about a set of equations describing your system that can be used to make some predictions, describe your data,…. However, in many cases it is only a cartoon of protein A interacting with protein B. I’m not saying that one approach is correct and the other is wrong. But I think this is why there are these disagreements you mentioned.

        Your post touches many very important issues and is very helpful. Not that many people are willing to share their experience, so thank you again for describing your journey!

      • I found the interview process very helpful, and I think that it can be a constructive learning experience even if one does not get offers. I absolutely agree with the tradeoff that Sergey describes; in my case, I see gaining the learning experience and meeting colleagues as being the more valuable use of my time. I did not spend much more time on preparing tenure track applications beyond the time I needed to spend anyway for organizing my research priorities, funding applications and papers.

  8. Pingback: A drop in the ocean known as “job market” | Genomicrocosm | Alexander Suh

  9. Hi Sergey, found your post through Reddit!

    It’s indeed depressing that Science paper is so important, but at least it’s a clear goal. I haven’t factored this in when I did my “cumulative impact factor” analysis back in the day

    So basically, if your theory is true, not just a talent and some good work, but the pedigree is really important. One should land a postdoc in a well-funded lab with a well-known PI if they ever hope to get a R01 faculty position, as that’s pretty much a prerequisite for getting a CNS paper, isn’t it? Good to know.

    I wish somebody wrote a meta-paper on the job market situation, analyzing some data of this kind. It’s really a pity that most of the data is either hidden from the public eye (short-lists); or not easily googlable (interviews). It makes your analysis even more unique and interesting. Thanks!

    • Thanks Senya, I think it is possible to get some interesting data by (manually) collecting CVs of recently hired junior faculties [Would need two-three undergrads for a month to do that]. Having long and short lists would be awesome, but it’s pretty much impossible, unfortunately.

  10. Pingback: The strange myopia of tenure | Science Dilemmas

  11. Pingback: Friday links: an evolutionary biologist’s apology, plankton art, and more | Dynamic Ecology

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s