Artificial Intelligence is no match for natural stupidity

Google designed an algorithm to solve its staffing problems. The algorithm is designed to help the organization identify employees most likely to quit. The Wall Street Journal put it this way:

The Internet search giant recently began crunching data from employee reviews and promotion and pay histories in a mathematical formula Google says can identify which of its 20,000 employees are most likely to quit.

Yeah. The folks who brought us the best and cool search engine, will now calculate and determine whom they might risk to loose. One of the boffins at Google, Lazlo Bock, believes the algorithm can get "in people's heads" and determine they will leave before they will know it. Though I can't comment on using an algortithm to predict who will quit, I am sceptical about the algorithm mentality already ruling the other side of keeping talent; I mean the one that brings talent in, or recruits new DNA into Big Brother.

I worked with Google recruiters in the past, they courted us by times to help them finding country managers for operations in emerging markets with fast growing economies;  but also with poor information infrastructure and legislation, and in some cases, markets that function in politically and culturally closed societies.

The power Google holds to search information on the web, and also on its own corporate networks, is immense. It permeates all aspects of its business, including talent acquisition. To my experience, the power of Google Corp. and its associated status leads by times to an attitude that is self defeating because the company looses a sense of realism most commonly closely linked to the admirable trait of humbleness.

It therefore fails to recognize talent outside predefined criteria of its search algorithms; criteria it confidently ascribes the power of dogma. While such criteria are quite valid in the First World, they are practically inapplicable to the type of emerging markets described above. The type of these c"hiseled in stone" commandments, such as Ivy League education, top university scores, and previous employment, can lead to the wrong candidates being hired, and the right ones not to be considered.

The internal recruitment managers Google apply similar algorithms to find candidates all over the globe, and of course, by times get stuck because the right professionals simply do not show up on their screens. This for various reasons, the most common one because the majority of professionals have degrees not from the "right" universities, do not have the "right" work experience, or not the "right" seniority.

For instance, the son of a poor factory worker in an ex-communist country who build the most successful portal in that country is not considered a candidate because he got his MBA from a UK University considered to be not "as good" as Europe's top universities. Hey, he has a local engineering degree, saved all he could get to receive a graduate education in the UK, to return and create upon his return the most valuable internet property and business in his country. The algorithm mentality completely missed a relevant candidate.

Furthermore, there is no algorithm providing a way to distinguish between false and correct information. I remember an "ideal" candidate to be whisked away to London to undergo the last interviews for a country management position in a North Africa. Only during the final due diligence of the candidate's education and career history it became clear he faked generously parts of his CV, prompting a colleague to swallow his "I told you so".

There is no algorithm that can capture the cultural finesse required to separate the wheat from the chaff. This issue won't be much of a deal breaker in the US or Western Europe, just don't extend it over-zealously to all markets and countries. A dose of some good, old common-sense modesty is sometimes the best antidote to algorithmitis.

As for an algorithm to identify people before they resign, must somehow be able, as the Google boffin Bock claims, know the thoughts of employees before they do. Next year we might get a full-blown psychoanalytic algorithm as result.

It's a brave new world, Dave.

Expert Knowledge Accessible to Anyone ?

People around Stephen Wolfram, the British Mathematican known for his famous software tool Mathematica, showed off a program taking Google a step further. The software aims to answer questions directly, the one "how many dwarfs surrounded snowy white" should give the answer 7, rather than a list of URL's like the usual search engines provide.

Even more complex questions, going beyond simple facts, should be in reach of the software, like a country GDP, the plot of a mathematical function, or the number of hairs still standing on my head. More about the program called "Wolfram Alpha" here.

If this software will fulfill even some of its promises, expect recruiting blogs and community sites to rave, contemplate and go bonkers about its implications for research - including this blog. Only I won't go on predicting the doom of Executive Search though.

Mirror mirror on the wall, find me the best.......

Search Execution

This post is a continuation of the entry I posted last week. When the business, and the need of a client is understood, the right search strategy needs to be defined and executed. The main question arising is where to look for potential candidates. Potential candidates are usually unsuspecting they are being targeted by search consultants, simply because they are not on the market looking for a job. They are mostly happy where they are, and usually satisfied because they build a trackrecord of success at their current employer.

Enter the research function. Within the search firm, researchers assist consultants, they organize the operational aspects of the search process, and all related logistics until the uccessful completion of a project. Depending on the qualification and seniority of a researcher, he or she will also be the initial contact with all or some of the candidates.

All Search companies should have a strong research function. Those that won't spend on research are typically firms that work on success basis, or companies where the expenditure on researchers does not have the same priority as getting candidate "CV's" to a client fast. Independent consultants must execute part of the research themselves, or will to some extent have to outsource part of it. But no-one in this business will succeed if he or she does not master the basic elements of search execution.

A significant part of the search activity is passive, and will cover a thorough query of the search firms' proprietary database, including web 2.0 internet sources (a dimension the more traditional firms are currently neglecting). The organization of this aspect of the search activity is crucial to the time it will take to comprehensively cover the total universe of candidates, in order to bring to a client the best available professionals. The other, active part of the search activity consists of targeted and direct identification, whereby researchers call companies and harvest the names, phone numbers and email addresses of people in specific positions, all which in turn are added to the database. The last source of names, effective especially for difficult niche searches, comes from the referrals provided by industry insiders. Search consultants, conducting hundreds of interviews a year, get to know quickly "who-is-who" in particular sectors and can cut fast and efficiently to the heart of the matter with phone calls to the insiders they know. The goal of all this labor intensive activity is to create what in search parlance is called "the longlist", a list that for most projects goes over 100 names.

One side note: an activity not directly related to any search project, but one that goes on continiously, is the daily data entry of names into the database, usually delegated by researchers to temps or interns. This information is found in business newspapers, magazines and journals, who are read routinely by consultants and researchers. Names of executives changing jobs, speaking on industry conventions, companies announcing to invest, etc.... are identified and entered into the companies database, or the relevant records related to the executives and their employers are updated. This includes of course the CV's of people on the market send in, because they either seek a job or a change in position. A proprietary database grows from the information collection entered during a search project, but also from the on-going low profile and low intensity data entry activity of temps and trainees.

A well-oiled, dynamic and efficient search execution will bring a long list forward with names that are effectively relevant to a search project. A researcher could likely not be familiar with a particular sector, for example, Technology, and more specific, the role of R&D in a Software Development Company. In that case the due-diligence and pre-research of the Consultant will determine the quality of the research and data provided during the project. Going more specifically, do we talk in this particular case about Embedded Software, Middleware, ERP Application software, or a subset or combination of these... ? It is clear if one does not like research, lacks a passion for learning, and is uncomfortable with unrelated, eclectic sets of information, one should avoid the Search field. As is the case with some of the larger search firms, every sector has a "practice group leader", who are usually senior consultants with a specific sector experience: technology, private equity, professional advisory services..... and so on. One of their main functions is to assist others that lack the sector know-how and experience. helping them up to speed and participating in the project when required. Though I am not a guru about Media & Entertainment or Manufacturing, I can always draw on the experience of collegues who know the particular segment inside-out, should my own personal research not suffice.

The succesful execution of a search is the result of a balanced cooperation between Researcher and Consultant. In most firms the blueprint of the search methodology is quite similar. What differs, however, is its application. Good consultants and researchers are adaptable and versatile, and can cope with most constraints. A constraint can be the lack of search experience in a sector, the narrow niche focus of a project - "find us a General Manager for an Industrial Diary Farm located in the middle of nowhere in Ukraine", or the lack of potential candidates - like in the case where the hiring company has a bad image and retains a Search consultant in an attempt to address the problem through the hire of new executives. One particular and unique constraint we once ran into, for a project in the Caucasus, was that we were paid very high fees to search for a series of senior positions for we received the job titles,and company names, but not the role description and responsibilities. The reason was that the key decision maker and buyer - a member of the local power elite - did not wish to interfere with the search as he literally told us it was our job to find out because he did not know, to keep everything absolutely confidential, and just bring him "the best people". In this case, our people on the ground put us in touch with various industry insiders, which allowed to learn the specific challenges of each portfolio firm, their culture and strategic intent, which put us subsequently on track to find the expatriate talent willing to take on the role. The ultimate intent of our client was to professionalize and institutionalize, which became crystal clear from the first handshake and very start of the assignment. This is the reason why search firms refer searches and projects to one another, they will not jeopardise a client relationship and search assignment due to a lack of regional, on-the-ground resources and accompanying research ability.

After the majority of research is completed (though it always goes on for the period the project is running), the subsequent goal is to transform the longlist into a shortlist, through identification of which interested executives are most relevant to interview. Someone once told me search consultants go out and surprise, even "bother" people with a career opportunity. This implies that some will never be interested to pursue an opportunity, and that for a myriad of reasons. Usually those not interested are too committed to their current employers which does not always mean they cannot be carefully coaxed away towards the position one seeks to fill. Others feel over-qualified, some might have recently changed positions, or were promised a significant promotion, equity participation, or there are private, family reasons.

Executives on a long list can be approached through a combination of calls and email. A document drafted by a researcher, and written by the consultant, called "a brief" gives an overview without mentioning the hiring company, and describes in detail the position and required profile. The more senior the position, the more personal, discreet and sophisticated the approach needs to be. Executives of this caliber often want to talk first face-to-face, and determine if the role would really fit their capabilities and expectations.

The consultant will, based on a review of the CV and a phone conversation, decide if he wants to meet a particular executive for an assessment interview, and thus plan face-to-face meetings with candidates that are most relevant. Assessment interviews mainly cover the soft-skills side of the candidate, and to find out if he or she is the right personality with the fitting character and attitude to join the client organization. In Executive Search assignments, more than 50% of the hiring decision is made on this "behavioral fit". Assesment interviews will be covered soon, in a future blog posting.

One last, important remark: throughout the execution, the hallmark of an excellent consultant is that he or she will keep in close touch with the hiring organization, mostly to sharpen the focus of the search though the discussion of skill-sets, profiles and characters of encountered interested candidates.

Summarized, a search project starts (1) with the understanding of a clients business (2) the definition of the clients real need (3) the formulation of an effective search strategy consisting of targeting (4) the right positions and people through (5) passive - DB and Internet sources and (6) active - targeted identification and referral from industry insiders. This in turn leads to a comprehensive (7) long list from which the work starts to build a short list through (8) contacting one by one every identified professional leading to series of (9) interviews with potentially relevant candidates.