Balazs P. • Hi Hung - sorry to say but I see no worth in it.

The article is good and the quote is more than true that training recruiters on Boolean brings little ROI. Why? Becuase it is not about the operators. It is not about the technology. Boolean is not about the strings.

Boolean is all about logic, practice and routine. That can be hardly tought.

Kenexa brought BooleanBar with more-or-less the same features but beside killing my browser brough no any value to me.

Hung L. • Thanks Balazs, great comment.

I think most people in this group may feel the same - it's a Boolean automator plus ATS - and most here would probably not feel the need to use it. However, Scavado is addressing a different audience - those recruiters who are salespeople first, sourcers second. There are many recruiters out there who have not the time, patience or let's face it, the skill set to be Boolean sourcers, but will make their money by being great salespeople. That's Scavado's market, and it's a big slice.

Irina S. • Hung,

I have met and like the people behind these products. But I have to agree with Balazs. I actually think that products that automate search have better use in the hands of sourcers, than in the hands of recruiters who are salespeople first. The products make recruiters believe that this search is targeted and covers the top results. However this part of sourcing just *cannot* be automated. (Other aspects - such as sorting and parsing - can, and actually the best search automation products do have those other aspects covered as well.)

Automating search is like providing a "good recipe for a dinner" (for anyone, at any location) or like having a robot cook your dinner in your kitchen. Even if you are not cooking-inclined it's not a solution for most.

Gary C. • Totall agree with Irina/Balazs - no one tool can replace human thinking. This article from Glen Cathey (although he's referencing 'artifical/semantic' search tools vs Boolean & what Scavado does, his main point is to 'think' vs allowing the tools to do the work:


http://www.booleanblackbelt.com/2012/01/talent-sourcing-man-vs-aiblack-box-semantic-search/

Just my 2cents!

Gary

Martin L. • Yes it can be automated. You can also buy an all day breakfast in a tin.

Seems a little one dimensional to me to have a product with set strings. How many times have we all slightly tweeked a string to give a whole new load of candidates ?

If they have enough people within a particular market using the same strings in the same location then they will find and contact the same people. This is not good for the candidates, recruiters and certainly won't give the client the best people for the job.

As Gary an Glen say you aren't going to get that intelligence with any software that I have seen (yet).

So no surprise I'm with my esteemed friends above in that until a semantic product is developed that really works then the best searches are those that are written fresh by a human, bespoke to the requirement and with a little love !

Saying all that I haven't used Scavado, wish them well and would love to be proved wrong.

Maybe the guys at Scavado could give a trial to members of this group who would give the best feedback anywhere.

Glen C. • Fresh off a Twitter debate with Hung, I found this and had to weigh in. :-)

First - I have to tip my hat to the other heavyweights in this discussion - esteemed lady and gentlemen.

Like Balazs observed, searching/sourcing isn't about the operators, or the "glue" of queries.

Discovering and identifying potential talent via text & data is about information retrieval, which is the "is the area of study concerned with searching for documents, for information within documents, and for metadata about documents, as well as that of searching structured storage, relational databases, and the World Wide Web." (From http://en.wikipedia.org/wiki/Information_retrieval )

Going one step further - Human-Computer Information Retrieval was coined by Gary Marchionini whose main thesis is that HCIR "aims to empower people to explore large-scale information bases but demands that people also take responsibility for this control by expending cognitive and physical energy." I want to emphasize the part where he specifically notes that HCIR demands that people take responsibility for the process and expend cognitive energy. Trying to automate information retrieval is like trying to lose weight by taking diet pills and not exercising and not eating right.

Information retrieval is a process, not an outcome. A query is a formal statement of an information need (skills, experience, location, education, corporate pedigree, etc.) - as such, you cannot automate the translation of everyone's information needs into effective queries that will yield relevant results, as the information needs are almost infinitely varied from person to person, role to role, and company to company.

I totally get why companies make software to "automate" search - because there is a perception that sourcing is difficult. It's actually not that difficult, but that's an entirely different matter. People are always looking to buy things that will save them time and effort, but not all things sold actually work as sold, and I am not singling out any particular solution or vendor here - not my intention. However, let's not fail to recognize capitalism at work - people see a need, they make and sell a potential solution that may or may not actually solve someone's problem (e.g. diet pills). There is an inherent danger in someone believing that if they are out of shape that all they need to do is take a bunch of diet pills. But if you watch the commercials and read the advertising - you could believe just that.

Another point most people overlook is that the only way to judge the effectiveness of any particular search is to review the results. You cannot tell how good a given query will be simply by looking at the query. Analyzing the results is where the "magic" happens - that's when you can tell how effective the query was by the relevance of the results - how many meet your specific needs vs. the # of false positives.

Additionally, it's quite rare that I complete a search with a single query. Most often I take my best guess at a query, analyze the results, and modify the query appropriately to increase relevance and decrease false positives wherever possible. Again - this is a good illustration of why it's not effective to have a set of preconfigured search strings.

Beyond query development and refinement, the real "work" or sourcing is the intelligent and insightful analysis of the results and iterative query improvement to produce a larger quantity of more relevant results (which is possible). Data requires analysis to yield information and knowledge, which in turn can be used to make informed decisions. As Gary pointed out (thanks for the mention!) - applications don't analyze anything - they only move stuff around. :)

To pound another nail in the coffin, contemplate this and then assess any "automated" search solution: Every query both includes and excludes qualified and available results.

Thank you all for the discussion!

Chandra B. • For 15+ years eGrabber has been successfully providing automation solutions that increase the productivity of recruiting & sales professionals.

I believe Automation tools have a place in the world and for a LARGE set of audience they provide a fantastic ROI when solving certain classes of frequently occurring processes.

As an analogy...
I often hear gourmet chefs say how bad packaged food or fast food tastes. Even though gourmet cooks maybe right, the fact of the matter is outside food is a great option for a LARGE number of people, and at times even for the experts.

There are similar analogies when listening to conversation between expert sourcers on forums or in conferences. Experts have a different need, they tackle a different set of problems and have different value system. They simply have different goals.

In my opinion functions that are routinely done by the majority of experts in a similar manner can be automated to a large extent. Also, in most cases what experts do today, will be automated soon by innovators so that the majority of the people can benefit from those techniques.

In the past only experts knew hot take great photographs with cameras. People need to understand lighting, depth, film speed, shutter speed and so on.,
These days a high school kid with no knowledge of how camera work can take a photo as good as an experienced photographer did 10 years ago.
Innovation happened, reducing a value of the expert photographers for most organizations.

So it is a fact of technology evolution that what most experts end up doing, will end up being automated.

So the basic question being asked
"Can Boolean search .. be automated?"
My answer is YES, they can be automated, so that the solution is adequate for the average user, to the same extent as how the expertise of taking photos and cooking great meals has been automated.

Glen C. • It's funny how accounting, financial, ERP (SAP, PeopleSoft, Oracle, Microsoft, etc.), and BI (Essbase, SAS, Business Objects, Cognos, SSAS, etc.) software has not automated accountants, financial analysts, ERP analysts, and BI analysts out of jobs.

In fact, the types of software listed above have arguably created more jobs than they have eliminated, including a whole range of highly paid functional expert-level positions in nearly every large company.

Furthermore, data science, which leverages machine learning, which is based largely on the idea of automation, is also now creating more jobs than it is replacing/eliminating. http://jobs.aol.com/articles/2011/08/10/data-scientist-the-hottest-job-you-havent-heard-of/

I agree with Chandra that on a purely black-and-white level, Boolean search can be automated - that's a simple thing really.

One could argue that any site/software that doesn't force you to manually enter in ANDs, ORs, and NOTs is "automating" the Boolean aspect of query building.

I do want to point out, however, that my main argument is that the analysis and understanding of the hiring need in the first place, crafting appropriate queries (not just selecting from a list of preconstructed queries), analyzing results and adjusting searches and making appropriate decisions cannot be automated, IMHO. Ultimately, much of this debate comes down to how broadly or narrowly you define "Boolean search."

Bob - to your point, in my ideal model, I would use tools like eGrabber, SourcePoint, Daxtra, etc. to automate the extraction (think ETL) of human capital data from the Internet, resume databases, social networks, etc., into a central repository (think data warehouse) with robust query capability (Lucene, dtSearch...) as well as automated matching technology and analytics - for the best of all worlds.

Lori F. • You can understand why Scavado has taken some heat from bloggers and consultants that make their living by teaching Boolean search string techniques.

Please acknowledge that I don’t mean any disrespect to the many practitioners of Boolean Search.

As far as Scavado not being an effective tool for recruiters, I have to respectfully disagree. Most Recruiters choose the profession because they enjoy talking to people, not writing search strings.

I’ve been a recruiter for over 18 years. I wanted to love Boolean but here’s the truth. It kept me away from what my clients were paying me to do… Connect with people. I was spending so much time writing strings that I wasn’t bringing enough new prospects into the pipeline.

My recruiting firm has been using Scavado to source candidates for our client projects exclusively for the past 3 years. For those of you that don’t know this, Scavado wasn’t built with the intention of being a powerhouse software company. It was built for my firm so we could be more efficient. Clients started asking how we could find passive candidates with tough skill sets so quickly. We showed them Scavado. They asked if they could buy it, because they too had the same frustrations with Boolean.

I enjoyed reading your responses and found them all to be very interesting. It seems people devoted to Boolean try to defend it as either an unduplicateable art or science. But I think most hiring and recruiting managers would agree that they don’t care how many results are in your string or which search engine you are using to find your prospects. They want good candidates to talk to that match the specified criteria. I don’t care if there are three results or 300, if something good surfaces from a search and I have a new candidate, a new person to source a candidate from, someone to save for Q4, whatever. I have done my job.

Glen, to your point I rarely complete a search with a single query as well. When you really understand what you’re looking for the options are endless. It is effective to have a set of preconfigured search strings. With a simple keyword change, my results are totally different. Just like yours.

It doesn’t mean your way is better than mine. If a product produces relevant results for the person using it, then it works. It’s pretty simple. Something else to note, a number of our early adopters that continue to renew are Certified Internet Recruiters that are very comfortable writing their own strings.

Scavado and other sourcing tools are not for everyone, but neither is writing Boolean search strings.

Irina S. • That's a great discussion! :)

Lori, would you like to give the group an opportunity to try Scavado? The majority of members in this large group are Internet search beginners who are interested in effective sourcing, so this would be a great promotional opportunity for Scavado. I think we can all benefit from the experience, just like we did with theSocialCV trial a week ago.

Regarding ideal automated solution, I agree with Glen. (Acquiring human capital data from the web has a very difficult and interesting sub-task: matching different pieces of the same person's profile with each other.) I believe that theSocialCV and Talentbin have made some good progress towards "the ideal solution". By the way, Talentbin uses Lucene; theSocialCV has semantic search elements.

Glen C. • I am thoroughly enjoying this discussion, and I want to take the opportunity to clarify a few things.

I don't make a living teaching Boolean search. I have nothing to sell, I am not a consultant, and I have been with the same company for 15 years. I make a living ensuring that my clients get the best talent when they need it, often in a competitive scenario where 1 to 20+ other companies are also trying to do the same. The only way we are successful is through finding the best talent.

Developing a sound, effective and repeatable information retrieval process has enabled me and the 100's of recruiters I have trained to actually get on the phone faster and speak with more people each day. The ability to quickly find the right people facilitates and enables more interaction with more people.

The more people a person can find that have a high probability of not only being qualified for the opportunity being sourced for, but also interested in the opportunity, the more productive and effective they can be.

I don't defend sound information retrieval skills as an un-duplicable art or science - I've been quite successful at duplicating the science of it for 15 years, and have trained people with no prior experience to be able to demonstrate their ability to outperform multi-million dollar search/match software within a matter of months.

Quite certainly, there is no right/wrong here - there are always multiple paths to the same goal. However, as someone who is capable of not only understanding the limitations of search/match software solutions but also demonstrating them, I feel obligated to share my experience, opinions, and findings.

Jung K. • To Glen's point, time to find and time to submit are very important metrics to track OR deliver against given that it's a rat race! It all starts with information retrieval. Even if you have a super large ATS OR CRM to mine from, you are still searching for leads and hopefully those leads can convert to contacts where you can match them to opportunities.

I'm amazed at the reality that it is possible for me to find more of the talent I need even when sourcing for super tough requisitions where you don't exactly know where your talent will come from. For example, no one says I promise to deliver 5 ready to hire java developers from Monster, 3 ready to hire .Net Developers from CareerBuilder and 4 ready to hire OBIEE Developers from LinkedIn. Who talks like that? Anyone? I know we process it in our minds hoping a combination of different tools will yield the necessary pipeline to deliver results. Yes, you will get placements sourcing from Monster, CareerBuilder and LinkedIn. However, you do need to have a decent to respectable understanding of what your information need is in order to retrieve human capital data that satisfies the needs of the user.

Cathy O. • I'd also like to chime in that I'd be up for an opportunity to try it. That being said, even with it, I wouldn't want to stop honing more & more my sourcing skills and productivity. This is because I love both sourcing and recruiting, but I know there are recruiters out there who don't really enjoy sourcing but they just want to be talking/ selling opportunities to an audience. Watson proved himself smarter than a human on Jeopardy so.... who knows? :) At this point, it's hard for me to believe that a software can be smarter than folks who've contributed to this great discussion. However, I'd love to try it out. :)

Gopikrishnan R. • I prefer Boolean search not to be automated. Rather spending time on customizing the tool, we can customize our boolean and use them om search engines.

Balazs P. • Chandra and Lori made a very important statement: business does not care about strings, does not care about sourcing, research or Boolean. Business does not want to understand HOW a candidate is made behind the scenes just because sourcing looks too ‘magical’ to them.

While it is a great achievement that over the last 5-7 years sourcing has become more and more visible business does still not yet see WHAT they see. They not yet can differentiate between sourcer and sourcer and knowing only they will get ‘extra’ sourcing service from a recruiter makes them satisfied and relaxed but at this basic level.

This semi-blindness is what helps these string-generator companies today make advantage and great profit in their business (sorry to say!) and this brings the question what we all here should work on in the next couple of years.

We have to reach the point to be able to tell the story of sourcing but in a simple and easy-to-understand way. We have to be able to well articulate the main differentiators but avoid to talk much about AND-OR-NOT or deep-dive data analysis.

It is not the science and art of sourcing which will help business understand it and we should work on to translate ‘sourcing intelligence’ to ‘business intelligence’.

Going back to the restaurant metaphor: the great story of the fast food industry has already been re-told and refined (if not destroyed). And yes, many of us sometimes take these products, however, we all right now know that the quality is poor and not necessary healthy.

This clear view on quality will have come, I believe, in our most-loved world too. Let it begin now!

Glen C. • Similarly, businesses and consumers do not care about HOW software and hardware is developed (e.g. iPhone) - they just want it to be easy to use and to work properly. However, behind all good software and hardware, there is a product development team using sound and proven design, development and test methodology and best practices, much of which cannot be automated.

I completely understand the desire to automate, but just because some things or some portions of things can be automated, it does not necessarily mean they *should* be automated, or that automation will guarantee the best results.

And THAT is what businesses care about - the best possible results (candidates, products, etc.).

Therein lies the rub - all search strings, whether generated by a human or software, "work" in that they return results.

Unfortunately, many people really don't care about whether or not the searches actually returned all of the best people available to be found - they're just happy to have some results.

Cathy O. • And this reminds me of a previous topic Glen had posted on, finding candidates for a req vs finding the BEST candidates for a req.... On another note, I think Scavado & other such software works/ has users because the bridge between beginner sourcer and master sourcer who can work super efficiently and produce good submissions to a req is a VERY LARGE gap. I am not a beginner sourcer but not a master like many of you. Yet, I know there are others who are even less skilled than me or just plain have less (or no) interest in sourcing & those are the folks that automation tools target. There'll always be a market for such tools, just like there'll always be a market for fast food. As Balazc said, great sourcers need to sell ourselves even better and be very highly rewarded even, because what great sourcers do is super truly awesome! :))

Gary C. • Regardless of what tools are used or what strings are created by man, you still need a human to decipher the results & go back to using tools or create new Boolean strings to refine!!!!

Ruth-Dinaz S. • No machine shall ever 100% replace a human [at least not yet in sourcing :-)]. The super computer might be faster than the human brain but in sourcing it is not just all about automated algorithms. It is great to see what I like to call in my unsophisticated vocab "spider tools" developing in the market place and more power to them as they get better but such tools don't 100% replace the human mind.Sure you should use them if you can to assist, efficiencize, automate repetitive functions and even help to narrow the volume on that first pass if you are not a savvy Boolean syntaxer but use your brains too plse! Success lies in using what works for you and you and you (and each is different). People and tools can co-exist. So lets "Live and Let Live" folks :-)