Candidate Data Reveal Hiring Trends

Most analytics leaders I talk to describe hiring the same way. Too many applicants, too few who can do the work, and a process that feels slower every year.

The candidate side of that process is easier to evaluate than most people think. At Dataford, an interview preparation platform, we collect reports from candidates post-interview, asking which company they applied to, which role they applied for, how they landed their interview, how difficult it was, and whether they got an offer. We now have more than 800,000 such reports. This article looks at what they say about how analytics hiring works in practice and what hiring teams can learn from that.

A quick note on the data. All information we collect is reported by candidates who are a self-selected group. Candidates report difficulty on a scale from one to five, where three is average. I treat the numbers as directional signals, and I encourage you to do the same.

Referrals Still Win

The clearest pattern in the data we’ve collected is how much the entry channel affects a candidate’s chance of being hired. Candidates who are referred to a job by an employee receive an offer 60% of the time. For both candidates applying online and those who have been recruited, that figure is, surprisingly, the same – 45%. New graduates recruited on campus receive offers 51% of the time.

These gaps in the different entry channels to employment have held every year since 2019. While offer rates have fallen for every channel over that time, the preference for referred candidates has not. In 2026, 49% of referrals have received offers, while 37% of online applicants have.

It’s not surprising. Referred candidates are already filtered. Someone inside the company vouched for them, so the higher offer rate mixes that real advantage with selection. These data cannot separate the two.

For a hiring team, that distinction matters less than it seems. Either the referral itself helps, or referrals bring in stronger candidates. Both are good reasons to invest in referral programs and ask why the online funnel converts so much more poorly. If recruiter outreach converts no better than a cold application, that outreach may be screening for the wrong things.

Interview Difficulty

Candidates rate interview difficulty on our platforms, which enables us to compare companies. The median company sits at a three 3.0 difficulty out of 5. The hardest loops in our data belong to consulting firms and quant trading shops. McKinsey is at 3.78 across 1,000 reports. Among large tech companies, Google is the hardest at 3.35.

An interview a candidate perceives as difficult can still be a good experience. Some of the toughest loops also have the highest share of positive experiences. For Bain, 72% of candidates rate their interview as positive; for Jane Street, it’s 70%. However, Canonical, also on the most difficult list, sits at 12%. The most likely reason is how long the process is. Glassdoor puts Canonical’s average time to hire at about 53 days across more than 1,500 interview reports, and every applicant completes a long written interview and an aptitude test before meeting anyone, followed by several rounds of interviews.

My read is that candidates accept a hard interview when it feels relevant and well-run. What they resent is an interview that feels random or disconnected from the job. For analytics roles, a good loop usually means an SQL exercise that looks like real work, a business case tied to the company’s own product, and clear communication about next steps.

Same Title, Different Pay

We scored 27 companies for data analyst roles on pay, work-life balance, and interview ease. Leveled total pay ranged from $79,464 to $248,400, with a median of $127,550. That is more than a threefold spread for the same title.

That spread shapes who applies where. A candidate who can pass a more difficult interview at a higher-paying company will often do exactly that. Teams that pay close to the median have an advantage they rarely use: their process. Walmart, for example, pays about $133,000 for an analyst role, runs one of the easiest data analyst loops on our list, and ranks sixth overall because of that combination.

If you cannot compete on pay, compete on process. A shorter loop, faster feedback, and an interview that tests the actual job are things candidates notice and talk about.

The Disappearing Job

About 7% of the interview write-ups we analyzed contain a concrete complaint. Ghosting is the most common one at nearly every company, followed by rudeness and disorganization. Slowness barely registers. The highest rate of complaints about a slow process at any company is 3%.

Candidates will sit through a long process. What they complain about is a process that ends without an answer. For a hiring team, this is the cheapest problem to fix. A clear decision date and a short rejection note cost almost nothing.

AI Roles on the Rise

We also track how candidates prepare for their interviews. Between January and June 2026, we analyzed more than 231,000 reported study sessions for interviews from more than 200,000 candidates. “AI engineer” is now the second-most prepared for role on our platform, ahead of “data analyst,” “data engineer,” and “data scientist.”

“Data analyst” lost 4.7 points of share across the period, and “data scientist” lost 3.1. Newer titles, such as “forward deployed engineer” and “GenAI engineer,” went from near zero to about 1% of demand in a single quarter.

My read is that the analytics skill set is being pulled into AI work. For hiring teams, it means the strongest analytics candidates are also looking at AI titles. Job descriptions that show how the role touches model evaluation, experimentation, or AI products will compete better than a generic analyst posting.

My Recommendations

Start by putting a real number on your referral pipeline. If referrals convert far better than applications in your own funnel, that is a case for rewarding referrals and fixing the top of the online funnel.

Then audit the interview for relevance. Every exercise should map to something the person will do in the first six months. Candidates will put up with a difficult exercise when it clearly connects to the job.

Tell candidates what to expect and when they will hear back. A one-page description of the loop and a firm decision date remove most of what candidates complain about.

Finally, watch where your candidates are going. If AI titles are luring away your best applicants, adjust the role, the title, or the pitch before you lower the bar. The full referral analysis, with trends by year and by role, is in Dataford’s report on whether referrals help in 2026. The demand numbers come from the Dataford Job Market Report 2026.