Fred Reichheld introduced the customer-focused, Net Promoter Score (NPS) in his 2003 Harvard Business Review Article, “The One Number You Need to Grow” and we're certain the connection to today's candidate experience is a strong one.
Apparently asking your customers a direct question like: “How likely are you to recommend our company/product/service to your friends and colleagues?” is a pretty solid indicator of your brand’s relationship with its customers and, in fact, has real value in differentiating one company’s performance from another.
Net Promoter Score is not without controversy. Just how far you can take your NPS as an accurate reflection of ‘customer loyalty’ and, just how confident you are that it will predict future growth of your service/product/company is open to a pretty hot debate whose resolution will not be answered or added to here. Amazon is loaded with useful books fueling the NPS debate for those wanting to dig deeper
However, in recent years, the notion that extreme responses to a single basic question- defined as “Attractors” or "Promoters" versus “Detractors” has become very appealing to the Talent Acquisition community as well as our Marketing colleagues. NPS is easy to measure, convenient to work with and, with a few practical protocols in place, it should prove a reliable measure of your candidates’ experience month over month, year over year. With a stable metric in place, your hiring practices that touch candidates can be tested- A/B style or otherwise and with both Marketing and Finance onboard regarding their understanding of the data, your analysis and the insights you derive stand a good chance of driving change.
At TalentBoard, the non-profit that oversees the Candidate Experience Awards each year, we’ve been refining a variation of Net Promoter Score (NPS) to measure the candidates’ experience- and, with half a dozen years under our belt, we’ve aggregated, with the help of 100s of firms (and quite a few willing experts on survey design), detailed responses from what is rapidly approaching 1 million candidates responding to every conceivable touch-point on their journey. (Check out the 1-day CandE Symposium in Nashville, October 2 for much more)
How the CandE NPS is Calculated and Why it's meaningful for Candidate Experience
We simplified the traditional method of calculating NPS this way:
We ask candidates as part of a comprehensive, 60 question survey this question:
“Based on your experience as a candidate, how likely are you refer others to [company name]?”
The responses to the question are limited to 1, 2, 3 or 4 (traditionally, NPS offers a 0-10 scale). The response to ‘1’ includes the anchoring phrase: “I would go out of my way to dissuade others from applying”. The response to ‘4’ includes the anchoring phrase: “I would go out of my way to encourage others to apply”. The responses to ‘2’ and ‘3’ do not have anchors and are ignored as neutral when calculating NPS. The % responding ‘1’ to the NPS question (Detractors) is simply subtracted from the % responding ‘4’ (Attractors) and the resulting score can theoretically be anything from -100 to 100.
The idea is simple. The higher the positive score the more Attractors outweigh Detractors and the stronger the connection between specific practices and positive attitudes and behaviors of the candidates. And conversely... We've found the dispersion of NPS using only 4 responses covers 50% of the scale and has shown remarkable consistency year over year.
For example, when candidates respond to a 5-point Likert Scale asking for their “overall” assessment of their experience (where a ‘1’ is negative and a ‘5’ is positive), the resulting average of all responses is not useful in any way, BUT, ~80% of those candidates who respond ‘5’ on the Likert scale also respond to our NPS question indicating that “they would go out of their way to encourage others to apply”. Conversely, ~40% of those choosing to respond ‘1’ to the Likert scale intend to “go out of their way to dissuade others from applying.”
The advantage of switching to NPS as a baseline measure is being able to correlate specific practices to candidate attitudes that are more likely to change behaviors to re-apply, refer others or even buy or reject an employer’s products or services in the future- not to mention the degree they will influence others.
Over the last 3-4 years, the firms receiving the highest ratings have had CandE NPS scores ranging from 20-60. The largest proportion of employers participating in the Candidate Experience Awards tend to have scores between 0-10 and a small portion are negative.
Two Critical Adjustments to your NPS score are necessary: Status and Timing
The status of your candidates is the most likely point of error. What proportion of your candidates who responded to the NPS question knew they were offered a job? How many knew they were rejected? How many were they simply in the dark? The answer matters.
Companies that only ask their ‘Selected’ Candidates to respond i.e. those offered a position are likely to learn little or nothing from such restriction of range. The score will be high. Much can be learned from including the ‘Finalists’ – everyone who got to an interview stage, assuming they knew they wouldn’t be offered a position at the time they responded. Why not ask everyone?
We believe the best practice is consistently asking the NPS response of ALL candidates who have applied, regardless of their qualifications, within a few days of each candidate being informed of their final status. The question could, ideally, be sent out automatically three days after final disposition with a compelling message asking the candidate for feedback to help the firm improve. This request of a candidate for an answer to a single question, providing an additional option to expand on their answer if they choose, will generate significant response.
At the least, you can filter responses by status but it is easy to normalize the score of all responses to avoid unfairly elevating or depressing your score. Those who know they are hired will naturally rate their experience higher. Those not hired will have a depressed rating. It is essential to determine a standard ratio of Hired/Not Hired to adjust to- TalentBoard normalizes to 80% knowing they were not hired versus 20% knowing they were hired (whatever ratio you use, just be consistent for comparison).
We would expect 10-25% response rates for this approach. Some employers have achieved 50%+.
When communicating with candidates, we also suggest promising that their response will remain anonymous- that the only thing the employer knows about the respondent is whether they knew they were hired or not when they responded.
Timing can be an issue. The longer you wait to request a response to calculate NPS after dis-positioning candidates, the lower your rating- especially if candidates do not believe they have been informed. As you pass well beyond their expectations for having heard from you, well, good luck with that.
Best practice is to set expectations for when candidates will be informed and deliver on them. Be consistent.
The Baseline for Candidate Experience
The NPS metric, adapted to Candidate Experience, can offer a baseline for how an employer’s Brand relationship with its prospects changes as they become candidates. Without a baseline for your treatment of candidates, you will simply be guessing.
[see also: CXR Candidate Experience Insights]
Full Disclosure: Gerry Crispin is a co-founder of TalentBoard. He receives no compensation from this involvement. This is true for the other founders, Elaine Orler and Ed Newman and, for all the volunteers – practitioners, consultants or, vendors supporting this initiative.
Not a member? You're missing out.
After logging in, CareerXroads members gain access to the final remarks of each article and the collaboration taking place among participating colloquium members. Other benefits include a full library of CareerXroads research, meeting materials, thousands of peer reviews and engaging conversation with vetted professionals.
Log in below or contact us to see if your company qualifies for membership.