Our efforts are to assign “probabilities” to a given candidate (incoming resume) that he belongs to:
Function X → 0.9 probability
Function Y → 0.7 probability
Function Z → 0.5 probability
So that, his “FUNCTION-PROFILE” may appear as follows:
Mr. A.J. Mhatre / Function Profile
(Graph sketch shown with bars):
Sales → 0.9
Marketing → 0.7
After Sales Service → 0.6
Customer Support → 0.4
Finance → 0.1
That resume
The “keywords” highlighted by “matchmaker” in that resume
Any keywords that got “left-out” during highlighting (but which ought to have been highlighted).
This review can only be done by human experts.
Why was that keyword “left-out”?
(a) → HR/Anyone/Inder did not include it (because they did not consider it as being important/worth).
OR
(b) → That keyword did exist in “matchmaker” but somehow software missed it in the resume (e.g., wrong spelling?).
As far as (b) is concerned, I suppose you will find a way/method to ensure that this does not happen.
As far as (a) is concerned, it is a slow, long drawn-out effort. Over next 6/12 months, we should raise (by adding) the keywords in “Matchmaker” from current 13,000 to maybe 20,000 or so.
Once the keyword database in “Matchmaker” grows substantially, the probability of any keyword being left “unhighlighted” will sharply reduce.
One method by which we plan to add new keywords is by redesign of MEMBER SEGREGATION TOOL (as we discussed few days back). Here, below →
INDUSTRY
FUNCTION
(Add → Keywords justifying this selection … Keywords justifying this selection …
This modification to Member Segregation Tool will enable us to “capture & embed” into our knowledge-base the knowledge of a HUMAN EXPERT.
(SECOND METHOD)
In Module 1, Member-Search Engine, I believe you are planning to add a feature of “keyword-based search”.
If a consultant enters a keyword & there are no “results/records”, then that consultant will be forced to “CATEGORISE” that (supposedly new) keyword as:
Skill
Knowledge
Attitude
Attribute
Educational Qualification
etc.
(9 boxes indicated)
So, at night, this NEW keyword gets added to the “Matchmaker’s” Knowledge-Base & re-matches all the resumes in the database against this NEW keyword.
This was a bit of digression/recapitulation.
Let us get back to FUNCTION-PROFILING of an incoming resume.
See ANNEX: A.
Here, I have drawn three graphs, one each for functions:
SALES
MKTG
AFTER SALES SERVICE
For all of these, we took 80% of total occurrence as cutoff point.
Although cutoff in all cases is 80% occurrence, the actual number of keywords (that result/cause this 80%) will be different in all 3 cases.
(Sales) In one case, these may be 17
(Mktg) Second → 26
(After Sales Service) Third → 43
And it is with each of these nos. that we try to match the actual keywords…
…that we try to match the actual keywords contained in Mr. Mhatre’s resume, where we find:
- Sales → 15 words found matching the (17) words⇒ Probability = 15 / 17 = 0.9
- Mktg → 19 words found matching (26) words⇒ Probability = 19 / 26 = 0.73
- After Sales Service → 25 words found matching (43) words⇒ Probability = 25 / 43 = 0.6
And these figures are what we plot to construct Mr. Mhatre’s FUNCTION PROFILE as shown on p.1 of this note.
Now, at this stage, I wish to introduce the concept/feature of what I call:
PROFILE MODIFIERS
“Profile Modifiers” are those keywords that rarely/seldom occur in the STANDARD PROFILES which we develop for each function.
That means their “frequency of occurrence” is 1 or 2 or 5 out of total occurrence of (maybe) 3000/5000/10000.
These words are at the very bottom of the heap/queue (being rare).
In Annex A, these words would show up at the extreme right-hand side, in the areas marked RED.
For sake of simplicity, let us say that the bottom 10% of occurrence comprise such RARE keywords, whom we call PROFILE MODIFIERS.
Again, it is quite likely that,
Bottom 10% in function:
SALES = 69 keywords
MKTG = 94 keywords
AFTER-SALES = 106 keywords
Being at the bottom, we tend to disregard these as being NON-REPRESENTATIVE for that given “function,” although numerically these keywords form a MAJORITY (of course, taken together, not individually).
Question:
Why should we take any “notice” of such rarely-occurring words?
What could be their “significance” to the profile of a candidate?
Answer:
Significance/importance of such “RARE” keywords is that they MIGHT possibly describe a UNIQUE attribute of a person, which could be valuable to a Recruiter.
Examples:
GERMAN (Language)
Entymology (a subject of research)
X-Ray Metallography (a technique for analysing metals)
D.Sc. (an Educational Qualification)
.Net (a Skill)
Egypt (place of work)
As of now, we have no clue/idea as to how many (and which) of such RARE keywords exist in Matchmaker’s database of 13,000.
The issue of displaying such RARE keywords on PROFILER GRAPH
…is irrelevant, if we argue:
“If a recruiter or a consultant is on look-out for such an attribute in a candidate, all he/she needs to do is to enter that keyword in search engine, in the TEXT search, and if such a candidate exists, he will get caught/short-listed.”
But, from my own experience, I have reached a conclusion that most of the times, a recruiter or a consultant is not even clear as to what he is looking for in a resume!
This may sound funny but it’s true!
So, I feel strongly that, while displaying FUNCTION-PROFILE of a given candidate (resume), we should also display his PROFILE MODIFIER KEYWORDS.
This can be done by taking bottom 10% of the keywords in each function-profile (Sales/Mktg/etc. eg:) & then matching actual keywords found in a resume with these.
We may even find that a keyword (already highlighted) in a resume does NOT even figure in a profile!
Does not matter. It only proves that it is so “rare”, we did not even think/consider to include it.
Let us display all such RARE keywords (whether found in a profile or not) occurring in a person’s resume, in a small box at the bottom as follows:
(Graph above shows: Sales – 0.9, Mktg – 0.7, After Sales – 0.6).
