Optimizing Application Flows for Quality Inbound Candidates

For seven years, I led the team that managed global inbound applications for Google (> 4MM annually). Our team's mission was to take "millions of dreams and turn them into thousands of jobs to build hundreds of products to change billions of lives". I’m a firm believer in the effectiveness of an online or inbound first recruiting methodology because of the diversity and quality of hires you make and the cost effectiveness of the methods.

Building this strategy feels especially topical right now, as many of the conversations with talent leaders revolve around managing application volumes from many people desperate for their next role.

Introduction

Layoffs dominate the news, volumes of applications are at all time highs, trends’ data articles are showing the huge shifts in applicant volumes per jobs. Talent teams are adapting to being much smaller and doing more with less, and bandwidth is being stretched by the volumes of applications they are seeing. This shift increases recruiting costs and creates extra work that makes hiring the right talent so much harder.

The status quo

Typically, recruiting processes look like this—some better, of course, and many worse!

• The talent team posts a job, and they get a bunch of applications, employee referrals, and maybe a bit of sourcing too.
• The Recruiter reviews a sample of resumes until they have what feels like the right amount that look good enough to push them into the process, unless the role is now likely to be filled internally, in which case it’s likely no applications will be reviewed.
• The Recruiter will leave the role posted because it acts as a comfort blanket to hiring managers to see it out there, but the vast majority of applications will be unreviewed while the first batch is assessed and moved along.
• The unreviewed applicant pool will only get screened if the first tranche of candidates starts to drop out of the process through rejections.
• In the end, the recruiter will make a hire from the first candidate who makes it through the process, then close the role and bulk reject unreviewed applications.

This chart from an Ashby Analytics Report illustrates what I often see:

The problem

In no particular order, this way of working will cause the following problems for a company:

• The noise of all the superfluous applications will cloud the talent pool and make it very difficult to hire the best candidate for the role.
• Screening all these applicants, the messages these candidates send to your teams, and any other processing required add cost to your hiring.
• Finally, there’s the unseen problem: working this way impacts your ability to hire a diverse workforce, statistically.

Over the years I have seen data that suggests only the first 10% of applications are often reviewed in this kind of process. Not all, because when we at Ashby look at aggregated data 50-60% of candidates get screened eventually, but when the thought process is based on “only screen enough”, then 10% is common and it works nicely for the math in this example!

Within your applicant pool, you will have different demographics represented, and many of those will make up less than 10% of your applicant pool.

Why is this an issue? By not screening the whole pool, you run into statistics. Statistics are caused by your actions within the talent pool that make your workforce less diverse and lower quality.

Let’s look at a hypothetical example. You get 100 applications on average per role. "Demographic group A" is 4% of the labor force for that job and 4% of your applicant pool.

You only review 10% of the applicant pool.

The statistical chance of at least one of the four applicants from "Demographic Group A" being in the first 10% reviewed is around 34%. So on average, only in one in three roles will any applicants from "Demographic Group A" have their resumes reviewed.

If only 25% of applicants are qualified, based on your requirements, this drops to a 10% chance that the qualified member of "Demographic A" will be screened.

To put it another way, only in one in ten roles at your company will a qualified candidate from "Demographic group A" be screened and have a chance of being hired. and they will still have to compete with other qualified applicants every time.

The same goes for the best candidate, regardless of the demographics they belong to. In this scenario of 100 applicants, by definition, there is only one best candidate from the 100. There’s only a 1 in 10 chance they get screened into the first 10%, and only in 1 in 10 roles will your company even consider the best applicant!

You might feel that 4% is a low number. Consider 13% of your applicants are coming from one demographic. Only in around 77% of your roles will you screen a candidate from a demographic that makes up 13% of your applicant pool if you only screen the first 10% of applicants.

So this model of only partially screening applicant pools drives a continuously reduced quality and diversity of hires. Even if you get the screened percentage up to the 80% mark, over time and at scale, you’ll be impacting the diversity and quality of your hiring.

Why does this persist? Well, part of the reason is that sales data models have for years been used in recruiting. You’ve all heard of the waterfall model in sales, where X input equates to Y output after all the sales processes are complete, right? If you put more in, then you get more out, and the output is revenue. Bang the new sales gong more times, folks!

You have a finite number of roles, so more applicants or pipelines become costly and may deteriorate the experience. The challenge for any company is to bring in the right amount of qualified pipeline for roles with the right mix of qualified applicants (diversity of pipeline) to create a pool that an unbiased, structured interview process can assess to bring you a selection of fantastic candidates at later stages to hire from.

Only screening 10% of your pipeline does not optimize this situation! It might seem efficient, but instead, you’ll either make a lower-quality hire, revisit a stale pool of applicants, or ultimately source more.

Data models and analytics don’t exist to tell you in real time when to stop your pipeline with confidence, so the role stays posted.

Finally, because there is a myth that "great candidates don’t apply for jobs," these pools of applicants are allowed to grow - and then when you do run the pass-through data on the inbound channel, the ratios look terrible. That myth was around when I started in recruiting and comes from the days of newspaper ads. The very greatest minds on this planet search online for information every day, and that includes browsing jobs in their field, just as I like to take a peek at the houses on the market near where I live, even though I love my home! This myth comes from a mindset without technology. With the spread of information and the accessibility of vast amounts of connections, this has been mostly debunked.

So if too many applications are not optimal, what do we need to do to reduce the workload without compromising the diversity and quality of the pipeline? How do we get a smaller number of applications, enabling the whole pool to be screened in an efficient way?

Not the solution

• There are some serious implications if talent teams react to this situation by not advertising roles. Implications that will lead to a lack of fairness and inequitable outcomes, as employee referrals and sourcing from well-known companies drive hiring from closed talent pools, leaving unknown talent with the aspirational goal of working for your company in the dark about opportunities. The same goes for teams that choose to cherry-pick through applicant pools with biases like past companies worked for, schools attended, etc.

• Source through the applicant pool for candidates who have worked for target companies, and then reject the rest. Whoever a candidate has worked for is not a qualification. You will miss out on great talent, you’ll impact the diversity of your hiring negatively, and you may run into risks around compliance.
• Source school names with a filter for the same reasons.
• A great way to think about what not to do would be to think about bias-free resume screening, where if you were going to screen resumes after certain information has been redacted to reduce bias, then the information that you would likely choose to redact is the information you shouldn’t be sourcing against your applicant pool by which to select.

The solution

This is my proposed solution in six parts.

Data

Build a pre-hiring report that shows past hiring data for similar hires, where the role was posted externally.

• How many applications from all sources did you receive?
• How many of those applications had an action taken that was NOT part of a bulk rejection process? Count the applications that were screened in or out during the pre-hire decision timeline.
• What was the source of these hires?
• Make an assessment of how many applications you’d need from any channel to make a hire.
• Make an assessment of how many applications you’d need from just inbound.
• Decide on the source mix of applications you want to target; in effect, are you going to spend resources sourcing for the role?
• Use Application volume x pass-through rate modeling to build a picture of Apps:Hire and the number of candidates you expect to reach each stage.
• What will your hiring strategy be? Will you open the role internally first? Having a consistent internal, fair, and equitable process creates huge cultural trust within your organization, reduces the total number of roles advertised externally, and therefore reduces costs.
• Suggested order of operations, optimized for cost, diversity, and quality: internal first, then, if there are no suitable candidates, inbound; outbound sourcing if required; and, as a last resort, agencies.

Here's an example Ashby Analytics Dashboard:

Job descriptions

Build clear and transparent job descriptions that use standard terms, are optimized for discoverability and make it easy for a candidate to select in or out if qualified.

• Include minimum requirements based on the skills and experience that will be directly required to be successful in the role and achieve the predicted results that were used to justify the role being open.
• Use an industry standard job title so that candidates can find the role easily and so that job matching tools can push the role to qualified pools of talent.
• Include the critical keywords for the role. Do not be lazy and say, for example, "Computer programming”. Instead, write a list of the languages your organization uses in the space the role will be within. These keywords will be used by candidates to search for the best-fit role, so they must be present.
• Try A/B testing different job titles. A study I saw once suggested that a job title of fewer than 19 characters gives the most gender diversity in applicant pools. I don’t know why, but that’s what the study found across tens of thousands of jobs. I wish I still had access to the source study!
• Through watching application sessions at FullStory, I saw that only 33% of applicants take note of the minimum qualifications for a job. Instead, 66% apply after only looking at the job title, so you still need more gating at the application stage to create a barrier to entry.

Transparency

Being transparent in your job descriptions further enables applicants to make informed decisions about applying.

• Put a closing date on the job posting, or even better, explain that the role will be taken down when a certain number of applicants have applied. All based on the data you produced in Section 1.
• Explain the interview process, time cost, and timelines so that applicants can decide if they are willing and able to commit to the application process.
• Be open and transparent about the salary offered. Use as tight a band as possible. Don’t just do it to be compliant with California. Do it because it’s the right thing to do everywhere.
• Add a few lines on “why not to apply” to enable applicants to make a more informed decision.

Systems

If your application system allows, use system controls to impact the volumes and reduce the duplicate, uncommitted and unaligned applicants. Ashby All-In-One supports all of the following methods.

• Use application limits and blocking rules that stop duplicate applications from the same human. Here’s a post about the application limits available in Ashby
• Ask open-ended role and required experience questions that require the relevant amount of experience to answer succinctly, and make these questions required.
• Cover all administrative questions like permission to work in a location, professional licensing, etc. in the application form.
• Use application logic rules to create auto-reject criteria based on the answers to yes/no or multiple-choice questions in combination with the required free text answers being completed.

Method

As we’ve already discussed, your applications form a “talent pool, not a talent stream.”

• Treat the pool as a pool and screen every applicant.
• Those who are fully qualified for the role should be moved to the long list stage. In Ashby All-In-One, I create a secondary resume review stage that I move all qualified applicants into.
• Based on the data you produced in step 1, you may need to further shortlist the qualified applicant pool. To do this, use the "nice to have" skills stated on the job description as an objective set of criteria, and ideally, have a second set of eyes do this review.
• Continue down the nice-to-have list of skills in order of priority until you have the right number of applicants to start the structured interview process.
• The extra time taken to wait for the applicant pool to fill and to screen the best candidates twice will be worth it because all the candidates will move through the rest of the process at roughly the same pace, and your hiring manager will be wowed by the concentrated quality and diversity of the pool you have created. Time to fill and time to hire WILL reduce if you follow this method, not increase, I promise you, I've had to prove this multiple times!
• You’ll hire the best candidate for the job, not the first one past the post.

Continuous measurement

As with any change you should continuously measure the impact of the change to ensure it is still fit for purpose.

• Tactical measurements, while the role is still being hired, will give signals like when to source because the application quality has been too low compared to the benchmarks set in Section 1.
• Strategic measurements over a longer period could include:

• Measuring the quality of hire improvements
• The diversity of hires
• Productivity per recruiting resource
• Cost per hire
• Application to offer PTR
• Future spend on fewer LI Recruiter licenses... my particular favorite measurement, because less sourcing is required to fill your open roles.

Candidates

If you’re a candidate reading this, I hope you have luck in your job search soon. The methods above will help you, not hinder you, if you meet the requirements of a role. They will help you stand out for the jobs you have the experience, qualifications, and skills to excel in. When you find that matching job, please give the application your full attention.

• To stand out, try and explain what you did, how you did it, and the impact you had doing it. There’s an easy formula for that that you can use in the bullets of your resume. "I achieved X as measured by Y by doing Z." Credit goes to Laszlo Bock.
• Don’t apply for roles that you are not qualified for. You may run into application-limiting tools that block you from the roles you are qualified for.
• Your time is valuable too; don’t waste it on roles that are not a match.
• Read the minimum requirements for a role. Don’t apply based only on the job title.
• Tailor your application and resume to ensure you are showing evidence that you meet the minimum requirements stated in the job description. Often, companies are audited on their hires to ensure they are being objective in their hiring against stated criteria. This means they’ll only screen you in if you meet the minimums.
• If you don’t have a degree, look for the words "Degree or equivalent experience". From my experience, when you see this language, it means a degree is not essential but is instead often only stated because of US immigration rules around H1-B and other visas that audit and compare past job descriptions for the same role in the company as the visa request to ensure they require a degree "or equivalent experience".
• Standardize the job titles on your resume to align with industry standards, even if you have a fancy internal title.
• Be careful about only changing titles to align with the role you’re applying for. When your intro says "accomplished seller looking for a sales role" and your job title says "Customer Success Manager,", the inconsistency will stand out.
• Make sure your LinkedIn profile and your resume match for dates and titles. It matters; recruiters will look. ATSs are smart; even if you don’t have your LinkedIn URL on your resume, your profile on the tool will often connect to it through matching your resume content.
• If you have finished a role, put an end date on your resume. There’s nothing worse than being in an interview and having your integrity questioned because you finished your last role three months before you applied and didn’t state it on your resume.
• Fill out the questions asked in an application honestly. They will be verified during the interview process.

Making it all count

Following the process I outlined will not be a huge shift from how you work now, if nothing else you’ll write better job descriptions and make some more hires from inbound -- if you test this model fully and measure the impact, even on a sample set of roles, you’ll be surprised at the results; the happiness of your clients, the time you save and the amazing humans you’ll hire to change your company for the better.