Analysis and tips from my data science job search

The tedious, grinding work of sending out job apps


  • Overall, I had a 3% offer rate, and a ~10% reply rate.
  • Chances of getting a offer were actually fairly high (~30%) given a company that was willing to follow up on my application.
  • Wisconsin was the stand-out best state for companies reaching out to me about my app, and then later offering to hire me. (Also the salary there versus cost of living is quite high!)
  • Mid-March to mid-April 2018 had many job listings, and was also the most successful period of applications for me.
  • A simple story was critical. As part of this, putting my skills section at the top of my resume doubled the response rate I was getting from companies.
  • Indeed was my most useful job search tool. I list the best search terms below.


For 6 months, I looked for jobs, during which I sent out 158 job applications. In the end, I received 5 job offers. Successful, but stressful. Let’s just say I am glad the economy and unemployment rates have been the lowest in about 20 years!
Behind the scenes, my job search was really divided into two phases: the first was the spring job search leading to two job offers by graduation. Unfortunately, the one of those I decided to accept was later withdrawn. MERGH! Right, so back on the job search, the next time a bit more successfully with three offers in around a month.
And who was I, applying? In short: I was a domestic student (no visa sponsorship is a big plus in some areas), with an undergrad in hard science (neuroscience) and my master’s in business analytics. I had no corporate work experience. My previous research, work, education, and hobbies are extremely diverse, covering a huge range of topics.


I knew going into my job search, that things were going to be difficult.

  • No corporate work experience, but advanced technical skills
  • Unfamiliar with resumes, business lingo, and disliking of phone calls…
  • Rather young for many business analyst positions (age 22)
  • Complex background is difficult to follow
  • Minnesota jobs are often targeted at older, more experienced professionals
    • Also not much of a fit with finance/banking jobs, which many MN jobs are
  • A graduate business career center that wasn’t fully prepared to help a candidate like me (they were obviously confused by my background, for one). They are used to more experienced candidates and MBA students.


  • Being a domestic candidate with US Citizenship!
  • Being open to (indeed excited by) relocation possibilities
  • A very diverse background which meant I had experiences linked to a variety of industries (overall a negative because of it’s complexity, but sometimes a positive)
  • Excellent job market


Process Progression Success Rate


The transition rate between email and phone screen is almost 100%, which isn’t very surprising as I am the gatekeeper there. Interestingly, the progression rate for in-person interview to offer was almost 100%. This pretty much means there are three stages: application, initial contact, and final process.
Thus (for me):

About 10% of applications result in a company contacting me

About 30% of initial contacts lead to an offer

Which is about a 3% success rate from initial apps to offers

Not included in these numbers are four outreach emails/calls triggered by the Carlson careers center sending out resume books (amusingly, those all came rather late in the summer, apparently I was last choice of the available domestic students…). I ignored all but one of those because I had ongoing offers at the time of them.

Temporal Distribution

Two phases to my search, divided by a European trip and my earlier offer falling through.

The only really interesting thing here is that some of my earlier applications came through rather months later. For example, I applied to Geico for a “Strategic Modeling Analyst” position on April 7th. I was contacted around then, but didn’t really see anything happen until July. I got an offer from them August 16th a few days after flying into DC, my last offer. That was a four month process!
I got my last three offers much more quickly than the first offers. Why? I don’t know, but I have some ideas. I was getting better at interviewing and interacting, undoubtedly. Was the job market was also getting less competitive with the most qualified new grads already having accepted offers? It could also have been that some roles were looking for more immediate hire, and weren’t interested in me when I applied, say, in January.
Applying anytime within the first two or three weeks didn’t seem to make much difference. LinkedIn claims that ‘in the first 3 days’ is better, and probably for their over-applied postings it is, but for other postings this didn’t seem to be the case. Companies usually started reaching out about three to four weeks after the initial posting, it seemed.
Overall, my most successful application period was mid-March to mid-April. I had a good reply rate in summer as well, but it appeared to me that there were much fewer new postings at that time for positions that suited me (versus the spring).

Geographic Distribution

Excluding four applications to overseas positions (all of which were never heard from again), this is where I applied to:

My job apps by state. Clearly a Minnesota bias, eh?    Gray-shaded = no apps; white-shaded = a few apps

My job search was focused on Minnesota. My secondary search area was DC metro area based off enjoying the area and there being lots of opportunities. And beyond that, any postings that piqued by curiosity I also applied to.
But the real question here is, were any states more successful for me?

Best States Ranked by Reply Percent:































*for all states where 5 or more applications were made
Of course, with such small numbers, statistically there’s quite a lot that could be random effects. I am also looking at ‘reply’ for this reason not ‘offers’, as with offers it becomes even more stochastic. But here are the conclusions I draw:

  • Wisconsin is clearly better than the rest. (I am deeply offended by this as a Minnesotan)
  • The D.C. metro area (which are the MD/VA jobs) is also relatively successful, especially consider it’s remote from me and businesses prefer local candidates generally
  • The Midwest has failed me here overall! Even though MN is home, I got only one offer here (which offer was later withdrawn by a company reorganization, so I am not sure it even counts!). Illinois is downright terrible in reply rate. I expected better!

A Simple Story

My greatest challenge, I personally believe was a confusing background (and therefore confusing resume, interview, etc.) coupled with a lack of corporate work experience. Many of my classmates with a few years of experience, even if that experience was only doing Excel work of the sort I could have done since high school, or less, had an easier time finding jobs. A consistent career trajectory was also very valuable.
I learned to make my background much more consistent by talking about my life long interest in ‘decision science.’ This interest led to my neuroscience background (and classics too) then off a bit into business analytics to focus more on the application of neural nets and decision sciences to real world problems (solving real world problems being something neuroscience won’t be doing much of for many years yet). This short story was completely true, but it leaves out quite a lot of my experiences and knowledge, and it took me a while to figure out which parts to emphasize and which to exclude.
I also did something else, abandoning the Carlson school resume format (which didn’t suit me, being more for people with years of experience to emphasize). Instead I did this:

Note how I start with skills right away in my resume. I did this in early March, and soon after the reply rate picked up significantly. I can’t prove causality, but it certainly is worth noting.

Putting skills at the top of my resume was meant to make it clear what I could do. Before, they had to sort through my less-related education and experiences. They could have sorted through those to see I was qualified, but these being HR people, they didn’t have much spare processing power. So skills went to the top (at my own initiative, and against counselor advice because it was ‘unconventional’). Before doing this, I had had an initial response rate of 7% and after it doubled to 14%.
In conclusion here, have a good (and simple!) story. Also, be willing to ignore advice from people with different experiences. What many people did wouldn’t have worked as well for me, and vice-versa.
Particularly advice from my parents, which was last relevant, by all accounts, sometime in the Paleolithic period. Also, don’t be like my parents and bemoan that I will never get a job. Really, doesn’t help. They were also very leery about me turning down my first offer, but in the end it was clear that doing so was indeed a good decision.

Job Postings

I, unfortunately, did not log the source of the job posting in my record of my job applications. That said, my memory is good enough to recall for many of the positions, and for most of those that replied to me.
Here’s what I remember:

Zero replies from LinkedIn postings

Most replies came from job postings found on Indeed

Probably the highest percentage reply rate came from postings found by searching directly on company’s career listing pages.

Indeed became my main source of job postings. It has tons of posts, and its search features made it fairly easy to narrow down to jobs that would fit someone of my experience and expectations. Indeed directed me to four of the five companies that gave me offers.
On the other hand, like I said, my highest reply rate came from corporate career pages. One offer came to me this way, by naively dropping in to the company site and checking for roles. The reason was fairly obvious: I was outright told sometimes that the postings there were, for various reasons, going up first for more of an internal crowd. Thus, much less competition. However, slogging through corporate career pages is very inefficient. So usually what I did was for companies whom I had already applied (via finding a posting on Indeed), I would check for additional opportunities at that time, and again later, on that company’s site.

My most effective search terms were:

  • “data scientist” and selecting ‘entry-level’ or similar on Indeed
  • “python analyst” – for more technical analyst jobs
    • ‘python’ was the most searchable and relevant programming language
  • “statistics” was sometimes a useful search
  • “data” as a general search term in small markets

I used more specific search terms in bigger markets (ie D.C. metro) and less specific search terms in places like Wisconsin and on corporate career pages. Just searching ‘data’ can be surprisingly effective. Yes, it gives tons of unrelated results, but usually the unrelated results can be skimmed over quickly.

On ‘fit’ of posting

Generally it seemed that offers came from postings that from the beginning seemed a better fit based off the job posting description. Many of the positions that I never heard back from were also positions that I was thinking ‘this is less likely to fit my background’ as I applied. I probably could have gotten a much higher reply/offer rate by sending out apps only to the best fitting positions.
That said, I would still probably do it this way again. Apps do take a lot of time in aggregate, but not that much time individually if done efficiently. If you can reasonably launch 10,000 ships, why not? Two of my best offers came from positions I hesitated about applying to, for example US Venture which seemed a boring industrial company from the outside. Also I got offers where two years of experience was desired, even though I didn’t really have two years of experience (at least not corporate analyst experience).
In general, I seemed to mesh better with mid-size, open-minded companies that were somewhere between industrial and tech companies. I had absolutely zero luck with graduation rotational programs, and I am not quite sure why. I also had no response from jobs overseas. That said, I only applied to four of the latter, and so perhaps if I had tried harder I could have had more.

Rant about online skill tests

I hate online tests and assessments for jobs, for the most part. I must have done well on some of them, because two of my offers came from companies having such tests. Yet, I did not enjoy them. Shout out here to Proctor & Gamble for almost making me blind with their endless visual pattern recognition section of comparing geometric figures.
Also I hate companies that send the assessment first. Yeah, no, I did that once but then ignored the other company that tried to do that. I am not going to spend four hours on tests for a company that will turn me down after a phone screen. Do the phone screen first. Although I didn’t have any major technical difficulties, online coding tests usually have one or two panic-inducing challenges that aren’t related much at all to the user’s abilities. And as for Epic, having an actual person watching me (through my webcam and computer actions) for the entire multi-hour test? Well, that’s just overkill, especially as they also had a test during my in-person interview as well!

Discussion of Offers and Final Thoughts

Of the final three offers I was considering, all three had good opportunities, seemed a good cultural fit, and promised challenges. Ultimately the deciding factors came down to money and relocation.
The $80,000 salary of my accepted offer in Appleton, Wisconsin is, according to an online calculator comparing cost of living, equal to $154,536 (via Numeo) or $237,560 (via in San Francisco. Clearly, although the exact estimates vary greatly, my salary in Wisconsin makes me far more money than I could make in the fabled lands of the West Coast.
Also my job title is “data analyst.” A low-level sounding term, but amusingly offering the same pay as a “data scientist” role offer. Job title and pay are not nearly as connected as I thought going in, and in fact job titles often seem quite random.
Although not my absolute highest monetary value offer, this offer in Appleton was probably the highest salary in reality. I will admit comparing salaries is important, if for no other reason than it is far easier to compare numbers than more vague things like ‘fit’ and ‘opportunities.’ In general, side benefits among companies are pretty similar.
To be honest, I probably would have taken any half-way decent offer in Minneapolis (indeed I did so take, except it was withdrawn). As much as a new place excites me, I love Minneapolis which is familiar and also simply, one of the best cities in the country especially in facets I care about (top in parks, fitness, good in cycling, acceptable cost of living, smart people, good opportunities of all kinds). That left me with Wisconsin and D.C., both areas, interestingly, with relatively little visa sponsorship. Wisconsin is a heathen wasteland, but still familiar enough, more similar to home than elsewhere. Also, wayyyyy easier to move to than the coasts. Fate has decided I shall move, and so I shall, to Wisconsin. For now, at least, we shall see, eh?
Overall, my job search was very much a learning experience. This is the first career-oriented search I have done. It was also challenging, as I was aiming high while also moving into a somewhat new space (at least from an external perspective). Next time, I hope, it will be much smoother!

My basic analysis and simplified data:

1 thought on “Analysis and tips from my data science job search”

  1. A post by the IEEE on LinkedIn data suggests there has been a small surplus of data science candidates in Minneapolis: “the biggest shortage of data science experts is in New York City (34,032), followed by the San Francisco Bay Area (31,798), and Los Angeles (12,251). There are a few surplus data scientists in Cleveland-Akron (1206), Minneapolis (832), Cincinnati (770) and a few other metro areas, but, reports LinkedIn, these surpluses “are relatively small and narrowing rapidly.” –

Leave a Comment

Your email address will not be published. Required fields are marked *