Video - Hiring for Startups - My talk at Talent Leaders Connect

Recently I was asked to speak at The Job Post event, Talent Leaders Connect.  I talked about startups, a little psychology and a hypothetical kitten kicking factory... no really!

 

Back to regular blogging soon!

The Mis-Match of Algorithmic Recruitment

It's the not so distant future.

A mobile app linked to a wrist mounted wearable wakes you, at precisely the right moment.  It monitors your sleep patterns and pulse rate and greets you each morning with a chipper "Go get 'em!".  You dress and get ready to leave the house, the fridge has emailed to remind you that you'll need to buy milk on your return.  You lock the door behind you with a swipe of your cell phone, keys are no more.  Outside, you step into a self driving car and take a different route to the usual commute - the car knew about the traffic before you did.  You arrive at work and boxes are moved into the previously vacant office next to yours.  You weren't aware of a new co-worker. There were no interviews. They were algorithmically selected from the passive talent pool.  Kept warm on a diet of Pinterest photos of the office and Youtube videos of kittens selected to be the most humanising for the Mega Corp you happen to work in...

As far as predictions of the future go the vision I offer above is hardly advanced.  The technology exists for the wearables, the Internet of Things and the self driving cars, it's just that last part that seems incongruent.

In the growing adoption of technology for HR departments seeking to differentiate their sourcing efforts, the idea of algorithmic matching is seen to be the magic bullet in the "War for Talent".  Beyond the clichéd war metaphors and gullibility of HR Tech buyers is the future of recruitment to be left to the robots?

Technology has made the discipline of talent acquisition better.  We've moved far beyond the data entry and green screen databases of a decade ago.  As a modern workforce migrates to online services so their digital footprint increases making them all the more easy for the new breed of sourcers to find.  Now the future, according to some, looks set to be the automated addition of new workers and a touted increase in the skill of selection.  I'm no Luddite but I can't help thinking this is a version of a technological utopianism whose primary supporters are those that seek to benefit financially from the adoption of the technology in question.

So many of the products available that claim to have solved matching are the same providers who don't recognise some of the fatal flaws that their products exacerbate. The primary example of this is the reliance on the quality of data on both sides necessary for a match.  The majority of matching systems are parsing CV's and then matching against a job description analysed in the same way.  This is exactly the limited key word matching that these systems say is so weak.  Even when other data are relied upon to beef up the input, suggestions of LinkedIn profiles and even LinkedIn endorsements are laughable. Especially in the case of unverifiable LinkedIn endorsements like mine for "Midwifery" and "Cheese Making".  Of course I'm totally brilliant at both of these things...

Even the more advanced of the matching algorithms that incorporate some elements of semantic search (context of search, location, intent, variation of words, synonyms, generalised and specialised queries, concept matching and natural language processing) are constrained both by the data the candidates provide and the job description or criteria the employer matches against.  Anyone who works in recruiting will be able to quickly see that both of these sources of data are flawed and subject to constant change.  Data in both these areas can be knowingly falsified, incomplete and always out of date.

This data is inherently flawed because people themselves are inherently flawed.  Candidates will always seek to portray themselves in the best light, hiring managers will always add some extra "nice to haves" or even make the work of two people into one mythical job description.  A matching algorithm is forced to make sense of too many moving parts and results will suffer.

In moving towards this style of recommendation the people in the processes are reduced to the status of commodities.  Subtle nuance is lost and the chance for innovation curtailed by inelastic parameters.  People are not a product.  When Amazon presents you with a book based on your buying preferences it has only to reckon with your fickle, transient tastes.  A book doesn't reject you because it feels it's too far to get to your house, or because the other books on the shelf don't feel your reputation is strong enough, a book doesn't want to work from is own home or have a counter offer from a series of rival readers...people do.

Recruiting is a constant stream of edge cases.  Whilst a matching engine might work for less complex roles at large numbers, it won't help you compete in winning that all important "War for Talent" you were so desperately spending your way out of.  The current level of technology is no match for the ability of a good recruiter.  This is not an indictment of the technology, it's an acknowledgement of the greater problem that exists in the institutionally flawed HR departments and Recruiting processes the world over.  Using a tool like this to gain another datapoint to inform decision making is a valid use - it's the shame of HR Tech that every new tool is paraded as "the answer".  If the industry could wean itself off it's obsession with the novel and shiny we might be able to tackle some of these issues at the root cause and realise that the skills we learnt whilst toiling at our green screens might not be entirely redundant.

What Developers Want - A Data-Driven Approach to Writing Engaging Adverts

When writing job adverts recruiters are often left to rely on a brief chat with the hiring manager.  They sometimes get input from one of the friendlier engineers and pair this with an old job description that has been slowly rotting on their  careers site for the past year.  The output of these less than ideal circumstances is a rehashing of the old job spec.  Some added promises of an exciting "culture" and an oblique reference to some new technology you may or may not get to use.  The advert is posted in the normal places and with little fanfare proceeds to garner a lacklustre response from candidates. A talent pool that is already bombarded with competing offers.

There must be a better way.  What if we could write a job description using the same words and phrases that our target audience are looking for?  If we could ask a large enough group of people what they are looking for then we could pull themes and even individual words from this dataset to create and advert that was engaging. Better yet, we wouldn't have to resort to the cliches and stock phrases from all the other job descriptions.

Coming by this dataset isn't easy, few people have the time to go out and interview the hundreds of prospective candidates needed to make it representative.  Even if an employer did this the data would likely be skewed by experimenter bias.  If only there was a way of reliably collecting this data from developers who felt free to say whatever they wanted.  Recently I discovered a way to do exactly this. Better yet the data was already captured for me.  

Hire my Friend is a new sourcing tool aiming to address the need for talent in the world of startups. Aiming to not expose that talent to unscrupulous recruiters or the volumes of spam they would receive on other sites.  Additionally it has some cool recommendation features, which made "endorsement" meaningful again.  I care more if a developer rates another developer highly than if the same assurance of expertise came from a colleague in sales, a school friend or their mum.  

On looking at the tool I noticed that candidate profiles, though anonymous and containing all the usual information, also asked one important question.  "What are you looking for?".  Suddenly I had impartial answers to that question from 13,000 (and growing) Engineers, Marketers and UX Designers.  After running a search for Ruby developers in London I had the data I needed, I pasted the answers into one long document and made that into a word cloud.  The larger the word the more frequently it occurs in the responses.

What Developers are actually looking for...

So what does this tell us?  Firstly that Hire my Friend's users are very much on target.  The majority of users are looking for work in small, startup teams.  It's the the details here that are more interesting for me.  I have always said that offering a job that is both rewarding and challenging is attractive, i.e. referring to actual problems to solve.  This is borne out by the answers given, the words problem, challenging, learning, solving and knowledge feature heavily.  The second biggest takeaway for me is the importance in stressing the "why" of the role you're hiring for.  Why is the work important? How will it impact the larger team and the rest of the company?  In describing the work we should ensure that we stress those elements that are "creative", "fascinating", "exciting" and "cool".

So given these answers how can we measure a job description against the data?  The same process can be used to evaluate our own job descriptions - here's mine

From the advert

For me the obvious difference here is between the active and the passive.  The job description has some of the same elements but still has some scope to be a better match. In a passive sentence, the subject of the sentence is acted upon rather than performing the action.  For a potential candidate this could mean that they are left with a sense of being used like a resource or that their individual importance in being downplayed.  What job seeker wants to be part of a massive swathe of hiring to become a cog in a machine? None I'd want to hire.  As William Zinsser says in his book On Writing Well, “active verbs push hard and passive verbs tug fitfully" a job advert should be a compelling call to action.

I'm going to use the Hire my Friend data to write different adverts and do my own A/B test.  It will be interesting to see if matching the word choice and elevation of individual over the companies own needs makes the difference I think it will.  I'll let you know how I get on.

Metrics that Matter

Firstly apologies to those of you that aren't quite as geeky about the numbers of recruitment as I am, I'll be back to ranting about the misuse of Pinterest for recruitment soon.  As I promised previously I wanted to give a little insight into those individual statistics that go to make up the metrics I use (or those I like to see) when recruiting.  Gathering this information isn't about producing a report simply to prove effort.  It is only the most unengaged stakeholder who can take solace in knowing that candidate and recruiters are somewhere in the building...  Gathering this seemingly disparate data points, in a consistent format (more on this later) is about creating a dataset that is alive and available to answer questions that may arise later... regardless of what those questions might be...

So what are the basics?  Those elements that you have to capture and whether that's in an ATS, a spreadsheet or typed up and popped in one of those old-timey filing cabinets.

Name, gender - All of your candidates will have a name, even if they have just one like a Brazilian footballer or Madonna they still have a name.  You should decide in advance on a format for writing these names capitalization, hyphenation etc this is to facilitate later use of names in mail merge or batch operations - candidates don't want to receive an email for "MAtthw BUCKLAND" so spell it right and you won't have to change 1000 name spellings at a later date.

Gender as a metric is of particular interest to me.  I've always worked in technical recruitment and it's an industry where females and transgendered people are under represented.  This metric can be combined with source to know which sources are productive for diversity goals and with the date ranges to know if and where candidates excel or fall down in your recruitment process.  This can facilitate later discussion and provide great evidence for changing processes later.

Role - the role the candidate applies for...this one really is basic to be able to slice numbers of total applicants by role, I hope everyone does at least this.  If not I guess they just tie CV's to the back of kittens and let them lose...

Gate Dates - Not Match.com for Farmers, this is the notation of the dates that a candidate moves through the hiring process.  Date of Application, Date of Phone Screen, Date of First On-site Interview all the way through to Date of Offer, Verbal Acceptance, Written Acceptance and Start Date.  GET ALL THE DATES!  So why track all these dates?  These date ranges can be used to answer a multitude of questions.  With values in these ranges reports can be compiled that show total length of process, drop-out ratios, expose bottlenecks in the process, expose waiting times and hold-ups, track notice periods... basically everything.  The date ranges and days elapsed are the bread and butter of recruitment reporting.  Do you currently know the average length of your interview process?  Does it vary a great deal?  Why is that? It's the interrogation of these dates that will give you those answers and perhaps when you have enough of an historical dataset predict time to hire of for future capacity planning... all for putting some dates in a spreadsheet or clicking those little calendar icons in your swanky new ATS!  Brilliant!

Source - Again a simple one, but it bears repeating, the source is how the candidate arrived in your recruitment process.  This should break down the source into broad categories that can tell at a glance what is a good source (a lot of quality candidates) a weak source (few candidates) or a bad source (lots of irrelevant candidates).  Example sources should differentiate between the "How" of the source too e.g. not just "LinkedIn" correct reporting should be "LinkedIn Search" and "LinkedIn Advert", this will enable you to distinguish between an active candidate application versus a directly sourced passive candidate.

Secondary Source - Some sources may require extra insight, you might need to know more for a later report.  If you have a primary source as "Event" this could be the particular Meetup, conference or pub you met them at.  A primary source of "Agency" might have the secondary source of the agency's name, for referrals it could be the refering employees name... remember they all have one...

Country of Residence -  I also like to track where a particular candidate is based this has multiple reasons, one might be for immigration purposes to highlight to internal teams where visa constraints may be an issue or delay a start date, a second reason could be to track individual sourcing efforts from a particular country... best of all most reports can include a lovely map showing where candidates came from...the prettiest metric :)

Contact Details - This should be the most obvious but still I see people finding value in the wrong things.  We all should know that a direct contact is better than a message delivered through a third party.  Simply put a telephone call or a direct email address are better than a LinkedIn Inmail.  If you only use LinkedIn to contact candidates and leaving it at that you're doing it wrong.

Last Employer - Want to know your pulling power?  Doing some competitor analysis?  Then you'll need to know where your candidates are currently working.

Recruiter - Who found the candidate and who is shepherding them through the process?  It's important that I'm not noting this to provide a productivity report for managerial consumption.  Unless all the members of the team are hiring for the same role in the same geography there is little to be gained from a direct comparison.  Raw numbers alone, stripped of context are not an aid.  They are a great example of one of the great flaws in gathering data - quantity isn't always preferable to quality.

Date of Last Contact - One of the consistent complaints and killers of candidate experience is the lack of timely feedback.  Even giving a candidate a short "no news yet" will pay dividends if you later wish to offer against a less communicative rival.  To overstate, if you track the last date you contacted a list of candidates you can very easily automate an email letting them know what's going on and when they'll get feedback.

Status - Decide on a glossary of terms that best fit your process, get the hiring managers involved in this process too.  Phone Screen, First Interview, Second Interview..etc.  Have as many of these as you feel you need.  Counting each of these each week will give you a very rapid view of the overall pipeline.  Hiring managers will love this, full on warm and fuzzy feelings.  Too often the work of the recruiter can look like a dark art - they go and stare at a screen and people magically appear for interviews - a weekly pipeline report just illustrating the numbers of potential candidates at each stage will calm even the most rabid of hiring manager.

There are more things to track of course and when real value can be derived from the collation of this data you'll find it quite addictive.  Best of all, when you start to move on from thinking the collection of data is just to describe the current status to instead thinking that you are creating a living, growing dataset that can be used to answer questions that haven't yet been thought of... you'll start to see why metrics really do matter.

The Itchy Security Blanket of Recruitment Metrics

The rise of more intuitive technology enabling the recruitment process has made for an interesting corollary - a rise in an organisation's ability to collect and report data connected to the recruitment process.  The increasing data driven programmatic approach to recruitment can do much to aid in the design and selection of a recruitment strategy.  Seemingly small changes can be tracked to measure their impact on the success or failure rates of a decision.

The growth in our ability to collect these metrics has been matched by a hunger within the stakeholder set as a whole.  Once a hiring manager has seen a report that gives seemingly scientific insight into the hiring process it will be almost impossible to revert to something which grants them less insight.  I'm not advocating that we take away metrics for these managers rather than we give them the access and supply the relevant context.  The greatest danger of data collection lies not in the information, but in its interpretation.  

So what metrics are appropriate to measure? What metrics can offer us certainty without falling into the the traps of selection or confirmation bias?  There are already a lot of hyperbolic blog posts like "The Top 10 Metrics You Must Have" or "7 Recruitment Metrics to Win" these miss the point.  The metrics of recruitment are best used for experimentation - tied to the continuous improvement of the team.  If you are producing metrics that will sit unopened in a spreadsheet to appease a hiring manager you are guilty of security blanket metrics.  Whilst you will feel all warm and fuzzy because you can prove that some *thing* is happening they will be of no real practical value, like butterflies pinned to a board underglass, nice to look at but not useful.

So whats the alternative?  When done correctly the term "metrics" is a misnomer.  The gathering of data around recruitment will give you a dataset which you can apply to provide insight into historical performance and to measure impact of the specific efficacy of projects the team undertakes.  In this way it's possible to see results in real time - does that new advert copy lead to more applications? You can see that! Which website is best to advertise on? You can test that! Did that rival companies announcement affect your response rate? You'll be able to see!  Did adding that photo of a cat to your website make it better?  Of course it did! You don't need metrics to tell you that!

What can't metrics do?  Predict the future.  In many of the articles I've read about recruitment metrics I've seen a large number of lofty claims about prediction.  All the while these claims are made without noting the limitations of the dataset we have access to.  It's the measurement of this dataset that will be the most effective use of business value not on fortune teller style inference of outcomes.  Statements like "we had 1000 applicants in 2013, so this year we will have 1500" are always going to be more wishful thinking than informed prediction.  Metrics can help in planning for the future but knowing the limitations of the basis of those predictions is key.  If we aren't aware of the limits of prediction we risk undoing the good that data can do and reaching for the crystal ball.  

In a future post I'll list the what and why of the metrics I like to measure.  Both for tracking team and individual performance within the team.  Hopefully you'll recognise it's a list high on building a dataset with experimentation in mind and low on fluffy feel goods and blame dodging.