Subscribe
The insider's edge on job search & hiring™

Glassdoor Salary Data: Worse than useless

In the June 5, 2018 Ask The Headhunter Newsletter we shake up salary negotiations and take down the Glassdoor myth.

Question

glassdoorI know you don’t like Glassdoor’s salary survey data and employer reviews, but what are we supposed to use to base our salary negotiations on? I’m talking about job seekers.

Nick’s Reply

If you’ve ever used Glassdoor salary data to help you negotiate a job offer, did you wonder whether you might have under-sold yourself? According to a report in the June 2018 Wired magazine by Rachel Nuwer, you might have left an additional 69% on the table. (Sorry, this edition of Wired is not yet available online.)

Wired tested samplings of self-reported Glassdoor salaries against Option Impact — a robust database of salaries reported directly by employers — and found glaring discrepancies. Option Impact is published by Advanced-HR for “elite users” — primarily venture capital firms and the tech companies they fund. These startups (and the investors behind them) can’t afford to make mistakes when competing for top talent, so they consent to share up-to-date salary information that Advanced-HR checks regularly.

In other words, unlike Glassdoor, which makes money off its data whether it’s accurate or not, Advanced-HR has a reputation to protect among VCs and the startups they run, and among the consultants and lawyers who serve them.

QA Engineer with 1 to 3 years’ experience

Option Impact: $101,955 (+43%)
Glassdoor: $71,004

Nuwer cites the example of a software engineer who got access to the database with his previous employer’s login. “Steve” turned down a handsome job offer of $180,000 that he says he would have gladly accepted — had he not learned from Option Impact that the reported market salary for that job was a lot higher. “His eventual starting salary: $205,000.”

The Glassdoor myth

I’m forever astonished at how easily people rationalize irrational behavior. Job seekers generally acknowledge that salary data and employer reviews on the popular Glassdoor website are biased and often phony. (See Can I trust Glassdoor reviews?) The salaries are questionable at best because they are self-reported. The web is rife with stories about HR managers and employers posting fake reviews to “balance” spiteful reviews from disgruntled employees.

Data Scientist with 4 to 6 years’ experience

Option Impact: $132,536 (+3%)
Glassdoor: $129,118

Yet I hear this all the time: “Well, I know all that, but you can still get a good idea about a company and what it pays by looking through all the information.”

No, you can’t.

Glassdoor admits it publishes, uh, crap

If you know some of the data are invalid but don’t know which, then it’s imprudent to trust any of it. Yet job seekers and employers peg their salary negotiations to anonymous Glassdoor “salary data” as if it’s a gold standard.

Glassdoor itself is clear in its Terms of Use that it doesn’t stand by anything posted by users or employers — that is, all its salary and company reviews:

“Because we do not control such Content, you understand and agree that: (1) we are not responsible for, and do not endorse, any such Content, including advertising and information about third-party products and services, job ads, or the employer, interview and salary-related information provided by other users; (2) we make no guarantees about the accuracy, currency, suitability, reliability or quality of the information in such Content; and (3) we assume no responsibility for unintended, objectionable, inaccurate, misleading, or unlawful Content made available by users, advertisers, and third parties.”

Sheesh. “Information” on Glassdoor is a myth. “Information” on Glassdoor is crap. What’s stupefying is that the company manages to survive and prosper by selling disclaimed “content” to suckers.

Glassdoor Salary Data: Worse than useless

The Wired report provides evidence suggesting Glassdoor’s salary data are worse than useless. The data are dangerous because they can actually cost you salary dollars when you decide how much to ask for. The job you’re negotiating for might be worth much more than the salary Glassdoor is “not responsible” for telling you it is.

Project Manager with 4 to 6 years’ experience

Option Impact: $137,000 (+66%)
Glassdoor: $82,403

Wired reports that the company behind Option Impact, Advanced-HR, doesn’t compile its salary data from employees who report it themselves — possibly fudging it. Advanced-HR gets it from the employers themselves.

“Companies share their employees’ anonymized salaries in exchange for access to the vault, which is searchable by job title, location, company size, revenue, and funding stage.”

Why would a company tell the truth about what it pays? Probably because Option Impact is an exclusive club and because these companies know venture capital (VC) firms rely on the data.

(Of course, not every job and industry is going to be in any salary database, including Option Impact, and all the general criticisms of survey data apply, including, Are we talking about the exact same jobs? I’m not suggesting Option Impact is the answer — just that it’s a fatal counter-example to Glassdoor’s swill pot of whatever anyone wants to pour into it. Advanced-HR demonstrates that there are other ways to do this.)

How can you get access to Option Impact?

Unless you’ve got access to some venture investment firm’s login, you’re not going to have access to the data that enabled Steve to get a 23% higher salary than he might have without Option Impact data at his finger tips.

Designer with 4 to 6 years’ experience

Option Impact: $126,125 (+69%)
Glassdoor: $74,591

So what’s my point, if you can’t get this data? It’s that if you trust your salary negotiations to salary data that you know is self-reported, unverified, untrusted, disclaimed and admittedly inaccurate (Thanks for the full disclosure, Glassdoor), you may be hurting yourself.

How can you get access to higher job offers?

Forget about getting your hands on valid salary data. It’s probably not going to happen. You’re not a VC or the CEO of a tech startup, and you probably can’t afford such exclusive insider data.

Instead, focus on the red meat of any job interview — be ready to show a hiring manager how you’re going to help drop additional profit to the bottom line if you get hired. Then you can ask for more money.

That’s a tall order, and there’s no short-cut. It’s why we’ve been talking about how to do it across hundreds of these Q&A columns. For example:

When a salary data vendor tells you it does not control the inaccurate, misleading information that it denies responsibility for, listen.

How do you know how much a job is worth? Does it really matter if you know how much you want? Do you use Glassdoor? How much would you pay for access to accurate salary data? What’s the secret to cracking the code of getting paid what a job is worth?

: :

 

 

 

20 Comments
  1. Crap. And here I sit with a master’s in a STEM field, not even making $13/hr.

    I gotta get a new job, really.

    • At the salaries alleged in this article, you’d have headhunters hounding every English-speaking university in North America looking for their CS / EE grads. Yet that’s not happening. I’m on the CS and Engineering jobs mailing lists for a handful of upper-tier programs, and they’re not exactly seeing that many inquiries. Many other ‘filtered’ venues (ie: ones where you can’t just buy a subscription and spam a thousand ‘jobs’) have virtually no postings for info-tech positions.

      Sounds like Wired unfortunately has fallen prey to the tech sector’s propaganda of trying to claim there’s a ‘shortage’ of workers when no such thing exists.

      • Mark: Headhunters don’t hound anyone. They look for low-hanging fruit – people who match keywords in job specs. I don’t see where Wired suggests there’s any shortage. In fact, Wired seems to be pointing out that the right candidates are so much in demand that employers are paying well for them.

        I’ll venture that the mistake you’re making is to rely on mailing lists and “subscriptions” to job spam services. They don’t work. Never have. After 12 years I stopped reporting on the surveys that showed in aggregate “job boards” deliver only about 10% of hires.

        There is no shortage of talent. The problem is the recruiting and hiring infrastructure. It doesn’t work. It cannot match people to jobs, with the result that HR and the DOL claim it must be because people don’t have the right skills. That’s been disproved.

        See https://www.asktheheadhunter.com/6827/employment-in-america-wtf-is-going-on

  2. Careful. Employers tend to exaggerate salaries, by adding in, amongst other things, holiday time, sick leave, other paid benefits such as free lunches, etc., when they come up with compensation figures. Glassdoor omits that.

    Given the sheer volume of applicants relative to few positions available in the contemporary tech sector, particularly in the Silicon Valley, the claimed higher salaries don’t really add up. The Glassdoor data is most likely closer to the truth than you think.

    We saw a lot of this in 1999/2000. Employees were even including the value of their stock options (most of which turned out to be worthless a year or two later) in their alleged compensation. That “$200k programmer” at the startup may very well be a $100k cash-compensated programmer and $100k of stock that vests 5 years from now.

    • Yeah, I’m wondering… do we really know that these data don’t suffer from the same self-reporting issues that Glassdoor does? I am more inclined to trust the salary numbers payroll processors put out, because it’s not self-reported at all, it’s aggregated data from actual clients paying actual salaries. Plus, employers have much more incentive to inflate their numbers than people leaving reviews on GlassDoor have to lie about their wages one way or another.

      I’m also skeptical that this particular salary data set is useful for anything *outside* tech. And even for tech outside major cities or that isn’t trying to compete with Google. I’m sure it breaks down salary by region/city, but still, the amount Facebook pays a remote engineer working in Kansas City is probably going to be different than what a small, local non-startup tech company is paying in the same city.

      And it definitely wouldn’t be helpful for me; knowing what Operations or HR folks in that data set isn’t going to help me because it’s all VCs and startups who can afford access to that data set, and those numbers are totally irrelevant for ops or HR working in, say, non-profits (even if we’re all *technically* in the same field, we’re not usually actually competing).

    • The people in my former company, in the heart of Silicon Valley, had no trouble finding new jobs when they were laid off or when they decided to clear out before being laid off. So I’m not sure where you get the plethora of applicants chasing a few jobs from. House prices here continue to soar to absurd levels also. At the lower level, almost every store you go into has help wanted signs posted.

      • Google, Facebook, etc., have applicant to hire ratios in excess of 1000:1. This is quite well known. Startups are inundated with applicants. Local candidates already employed in the sector may do better, but to join the sector as an outsider is notoriously difficult, especially if an American.

        • @Mark: So what does it tell us when the ratios are >1000:1? That there’s a flood of applicants and few jobs? Or that there’s something cruelly wrong with the application and matching process? I think the latter.

          As I pointed out in my column, Option Impact is subject to the same criticisms of any survey and I’m not suggesting it’s the answer to Glassdoor. My point is that Option Impact is an interesting data point that suggests there’s something very, very wrong with Glassdoor.

          Not that we need more evidence when Glassdoor itself publishes the kind of disclaimer it does in its Terms & Conditions!

          • Yeah a 1000:1 applicant to job ratio indicates that there’s few jobs relative to the number of applicants. Meaning that it is highly likely that large numbers of applications are being thrown out, sight unseen, not even considered by humans.

            A joke out there that is if Steve Jobs, or Steve Wozniak applied to the contemporary Apple company, they wouldn’t even get a job interview. The only problem is, its not a joke when talent of that level is excluded from employment because of the sheer glut of candidates relative to the jobs available.

            The discrepancies you cite between the Glassdoor figures and the “Option Impact” figures are perfectly explainable when you look at what an employee will report as ‘compensation’, versus the all-in cost of employing an employee, which is what an employer will typically report on an employment survey.

            Silicon Valley tech companies have little to no troubles recruiting everyone they need/want at Glassdoor-quoted compensation. The ‘shortage’ claims are only being manufactured to curry favour for more foreign worker visas, and to foment the training of more CS grads in an effort to drive down compensation.

            • @Mark,

              A question: Are you the @Mark from Disqus? There’s a Mark on Disqus that regularly mentions 1000:1 ratios at Silicon Valley employers.

        • Sure Google and Facebook have 1000:1 ratios. So do Harvard and MIT. That doesn’t mean that no one can get into college. Google and Facebook pay well but more importantly they have prestige. Silicon Valley companies, just like colleges, can be split into classes of perceived quality. The top ones recruit from the top colleges and other top companies. But the lower level companies have good well-paying jobs also.
          One of my former interns did get a job with Facebook, and a Google recruiter called me. So getting a job there is possible, but 1,000:1 for Google does not mean 1,000:1 for any job.
          That doesn’t mean there is a talent shortage, but there isn’t a talent glut either. (10 years ago, maybe.) People need to be realistic.

  3. I’ll be interested to read the Wired article. While Mark does have a point that employees often forget the non-cash benefits they have (my employer likes to remind us that we actually “earn” around $20,000 more than our annual income shows) there is also a serious problem trusting integrity of unverified, self-reported, potentially planted data. That is a big reason Auditors and Generally Accepted Audit Standards exist in the first place.

  4. Go seek out a good reference librarian – they also have access to information, including databases they have to pay for, that can give your information. Go to the reference desk and ask. This is what librarians do.

  5. Salary is just one aspect of a total compensation package. There are many other aspects such as health benefits, vacation, sick time, 401K matching, work/life balance, etc. that need to be taken int account before accepting a new position with a company.

  6. I don’t get why employers would be motivated to overstate salaries, since that is going to increase what candidates would expect to get paid.
    I do understand why Glassdoor numbers would be low, since I bet lots of people who go to that site are dissatisfied, which may be correlated with lower than average pay.
    Anyhow self-selected online salary data, like online poll results, aren’t worth the paper they’re printed on.

  7. Meaning that the salary numbers on GD can also be artificially high, I’m inferring.

  8. I have found that though it may not be accurate as you say… but when I go for interviews I usually would get to the point where the hiring manager would admit that they hit their budget cap… and I know it is accurate after the fact when I start chatting around with the powers to be or my fellow coworkers. Those numbers at least line up pretty well with numbers from salary.com. Is salary.com as bad as Glassdoor or am I stuck with that kind of salary grade from the companies I’m approaching?

    • According to them, “All of Salary.com’s data is proprietary and is based on employer-reported data. It does not contain data from individual site users, placement agencies, job postings, nor any other sources that would traditionally be characterized as “unreliable” by compensation or human resource professionals.”

      (source: https://swz.salary.com/docs/salwizhtmls/methodology.html)

      Note that salary is only *ONE* part of the compensation package. Other things include Job Title, vacation/PTO, training, equipment, benefits, sign-on bonuses, annual raises, golden parachutes, etc.

  9. The blogs used to be coming out about once a week yet the last one was about 17 days ago. Is Nick on a break or is the next one a really big entry?

    • I published a note in the June 5 edition of the newsletter that I’d be on vacation for two weeks. Next newsletter and column will be June 26. Thanks for noticing :-).

Leave a Reply