Sourcing: Alive, Kicking & Getting Loads Harder

My interest was piqued this week by an intriguing article written by Dr John Sullivan, entitled The End of Sourcing is Near….the Remaining Recruiting Challenge is Selling. It’s a well written and in places quite convincing article whose main arguments are that:

  • Now that most people have an on-line presence of some description, sourcing is easy
  • In the future, more people will have on-line presences and sourcing will get even easier
  • Soon, sourcing will be automated and at that point it will just be a commodity
  • Because of this, great recruiting should be defined by the quality of the sales pitch delivered to prospective candidates (or, in the words of a former colleague of mine – “hide and seek” has now become “kiss chase”).

Now, the ironic thing about all this is that sourcing has always been viewed by the recruiting community as “the easy part” of recruiting. It’s the reason why most recruiters start their career, having had zero training, as sourcers, earning their keep doing the “easy bit” until they’ve “progressed enough” (read: done enough placements) to move into the hallowed turf of Consultancy. And that’s not just true of agency recruiters either: joining an internal recruiting team at the bottom? Guess what you’ll be doing? Yep, sourcing.

That was, until recently. After years of being considered worthy only of the least experienced recruiters, sourcing has recently begun to emerge into a fully-blown, complex profession of its own. For instance, like many other established trades, we now have certification programs and on-line communities dedicated to sourcing. Sourcing has become a key topic at recruiting conferences as people trade ideas about best practice and new tools. New technology aimed purely at sourcing disrupts the landscape. The emergence of internationally-known, highly skilled and very well regarded sourcers (who I won’t name as I don’t them personally) has only happened because of a general recognition of the fact that sourcing is becoming a difficult, demanding activity.

And yet it seems, according to some – no sooner has sourcing established itself fully as an advanced skill-based profession, it is dead, over; yesterday’s problem.

Well, I’m sorry but I don’t agree. Sourcing is alive, kicking, and getting a whole load harder. And here is why:

1 More Data does not mean easier
If I gave you a book containing 50 mugshots and asked you to find the one who looks the most like John Smith, it’s not going to take you long, is it? So what about if I gave you a book with 200 mugshots? Or how about 10,000? You will most likely find a much better match in the book of 10,000, but is finding that match going to be easier, or harder?

I have been banging on about the opening up of the web for years. When I first started recruiting (as a sourcer, naturally) – for what was at the time a small recruiting company – we could access and search through maybe 250,000 CVs. Now that might seem like quite a lot, but when you consider that LinkedIn has 11 millions users in the UK alone, hindsight suggests that it really wasn’t very many at all. I can’t even guess the overall size of the candidate pool I can tap into now, but it’s in the many tens of millions. The sourcer in 2013 has a significant amount of data to play with.

It is hardly surprising that other commentators are latching on to the potential ramifications that the continued opening up of personal information has for the recruiting professional. There is no question that all this additional data enables us to do better sourcing, to find better matches to requirements. But the conclusion that having more data makes sourcing easier is very surprising indeed. More data means more searching. It means more time. It means more technique. More data makes sourcing harder!

It is more data, taking more time to process and more technique to navigate that has driven the professionalisation of the sourcing function. And as the amount of data increases, quality sourcing will get a lot, lot harder.

2 More Data means less quality
In the beginning there were databases. 10 years ago, the only data that recruiters had was held in either their own database or in an external, pay-to-search database. All of their sourcing activities started there. Whilst that meant that we had access to a lot less data than we have now, the data we did have was both in a usable format, and contained within a framework that allowed for easy searching.

CVs are the purest form of recruiting data because they are rich in all the information needed to make a judgement on a candidate’s suitability. Additionally, whilst I’m no fan of most CV database technology, even the most basic system normally offers a search-focussed interface and a variety of search options. The classic CV database – whether it be internally held or externally held, represents the perfect sourcing tool in terms of ease of use.

The reason that sourcing used to be considered the “easy” part of recruiting was because it was – it was the acquisition of the data that was difficult. Searching through that small amount of really good data using a tool designed specifically for that purpose was the easy bit.

The thing is though, classic CV databases have not been the cause of recent recruiting data growth: it is social data that has driven it (and to a lesser extent, platform-independent data such as that found on personal websites and blogs). And the thing with this data is that, from a recruiter’s perspective, some of it is useful, some of it is noisy, and a lot of it is just rubbish.

Let’s consider LinkedIn for a minute. Absolutely fundamental to the growth in recruiting data availability, LinkedIn changed the way that sourcing worked. It quickly turned professional information from largely private to emminently public. It opened out the kind of data that recruiters relied on for competitive advantage (and had spent years collecting) to everyone. LinkedIn made it possible for anyone to find anyone.

Sort of.

The reality is that the data LinkedIn offers – in the vast majority of cases – is nowhere near as rich and as nice as you would find in a stack of CVs. I don’t have the stats, but what percentage of profiles are fully complete? Whatever the answer is, it’s the vast minority. The vast majority of profiles are shells, containing piecemeal, basic information.

And yet LinkedIn is easily the best social platform in terms of data quality. Twitter bios are virtually useless. Facebook profiles are just as bad. To use either effectively for sourcing requires a level of skill and experience that most people will never develop. *Some* social profiles offer obvious links to user’s contact details, allowing an unskilled user to reach out to a potential candidate, but the overwhelming majority are – to an inexperienced sourcer – complete dead ends: partially usable indicators of shadow candidates.

The point I’m making here is that whilst we have a lot more data in 2013 than we had in 2005, the vast majority of that data is of a relatively low quality. We’ve gone from having a small amount of high quality data to absolutely loads of mostly low quality data. And that low quality data requires more experience, more skill and a lot more time to navigate and interpret.

It’s true that technology like (the really rather brilliant) Talentbin can be built to help us overcome some of the issues with data quality, but the truth is that as the amount of data grows and the number of platforms on which that data sits proliferates, the more diluted the quality becomes and the harder it becomes to use to any decent standard for recruiting.

Before we go on, let’s just summarize this:

Sourcing in 2005 Sourcing in 2013
  • Amount of available data to use in sourcing process was relatively small
  • Huge amounts of data available (and growing incessantly)
  • Quality of data available very rich
  • Quality of data much lower, generally made up of very basic information, lacking in important depth/details in most cases
  • Data mostly contained within databases specifically designed to be searched to extract relevant information
  • Data spread on multiple platforms, most of which not designed with effective search or extraction of relevant recruiting data in mind
  • Finding relevant matches a case of having the right data
  • Finding relevant matches a case of knowing how to navigate the data that everybody has access to

So, please, tell me: was sourcing easier then, or now? I heard one commentator this week say something along the lines of “now sourcing is solved, it’s all about selling.” Solved?!?! In a world where data is everywhere, sourcing has never been more unsolved. Social data is not the solution to the sourcing problem, it’s the cause of the sourcing problem!

Experienced, skilled and professional sourcers are the only solution.

3 Automation is not just a tech problem
I wasn’t going to talk about Dr Sullivan’s third major point, but doing that feels like copping out now. Hysteria about automation is not a new thing, but the idea that sourcing is something that can be conducted by software any time soon is – at best – very optimistic. And that’s because in sourcing’s case automation is not just a tech problem.

Even if the technology to parse and accurately interpret the meaning of a personal profile or CV to anything close to the level human accuracy existed (it doesn’t), the technical challenge represented by accessing enough data to make a 24/7, cross-platform, all-web sourcing bot possible is solvable by maybe one or two companies in the world. And even if THAT could be overcome, an even bigger issue is that whilst the growth in the availability of personal information is being driven by open data, that data is still owned by someone. And that someone isn’t you or me.

If you don’t believe that’s an issue, I double dare you to write a web crawler that scrapes profile information from LinkedIn. Go see how far you get.

Then there’s the issue of privacy. I’m not going there today, but how the web deals with privacy will have a very interesting impact on the sourcing and recruiting profession, and will throw a constants swerving curveball at any attempts to perfect automation (as will the behaviour of the users of social networks, which is completely undpredictable).

Glen Cathey – in his much better researched and far more articulate riposte to Dr Sullivan than this, entitled The End of Sourcing 1.0 is Near. Sourcing 2.0 Just Beginning – talks at length about the current state of AI technology and the various barriers that exist to full automation. His article is great and deserves your attention.

There is some fantastic technology emerging that makes it possible to do some great things with the data available on the web. But that technology is emerging to serve the sourcing profession, not to replace it. It is because sourcing has emerged as a stand-alone, serious profession that people are trying to solve some of the challenges that come with making sense of a proliferation of mostly rubbish data.

Dr Sullivan is definitely right about one thing: the ability to sell an opportunity to a potential candidate is an absolute defining quality of great recruiting. Knowing how to convert information into a prospective candidate, and how to convert a prospective candidate into a potential hire is probably the defining quality of a great recruiter. But you can sell an opportunity to the wrong people all day long if you like and the net result will never be great recruitment.

I am not surprised that sourcing has developed from the afterthought of the recruiting industry – the turf of the untrained novice – into a fully-blown profession in its own right. As those professionals find themselves using more complex tools to navigate greater volumes of much more fragmented and lower quality information to find the value-generating information, that profession will continue to grown in stature and complexity.

Selling has always been key to great recruiting. But one day, the sourcer will be king.

No comments yet.

Leave a Reply