EXCHANGE OF SELF-CREATED INTERVIEW QUESTIONS

Date: 19/01/08

Header: Rahul Swati Shalaka

Title: Exchange of self-created Interview Questions

We discussed this yesterday.

Primary Objective

To get this feature to act as a marketing tool/agent for IndraRecruiter. This feature should popularize/sell IndraRecruiter amongst recruiters.

Secondary Objectives

• To build-up a large database of IQ (Interview Questions) by harvesting the collective wisdom/knowledge of thousands of recruiters around the world.
• By getting thousands to "contribute" on an ongoing basis, manage to overtake competing job-portals and, forever, stay ahead of them.
• Change the "rules" of recruitment game and then play the game by "OUR" rules. We can never hope to overtake Monster/Naukri etc. if we play by "THEIR" rules!

Strategy

• Create a "COMMUNITY" of recruiters.
• Maybe, there are 30,000/40,000 recruiters (including our "consultant-level" professionals belonging to 5000 recruiting agencies).
• But they don't have a "forum" or a "platform" – whether physical or virtual.
• ERA represents only 140/150 recruiting companies—where only owners/MDs interact.
• As far as Corporate Recruiters are concerned, they are treated as "second-rate citizens" in NHRD (National HRD Network).
• There are 1/2 Lakh job-advts on Monster/Naukri etc, each carrying:
• Name of contact person (who receives email resumes)
• His/her email ID / Phone No.
• But do they talk to each other? Share — exchange any info? NO!
• Thru IQ Exchange feature, maybe we will succeed in creating such a VIRTUAL COMMUNITY.

• Create a PEER-TO-PEER network of these 30k/40k recruiters.
• On the web, there are many peer-to-peer networks. Some of these have millions of members. Prominent examples are music (file) sharing websites, where members exchange "music" files.
• Obviously there is some mechanism to know, which member has which music files.
• In our case, individual recruiters are creating their own IQs on any of the 86 IT skills / 43 Non-IT functions.
• NO recruiter could be an "Expert" on more than 3/4 skills/functions. So, she would have created IQs only for those 3/4 skills/functions.
• She may be quite willing to "donate" these self-created IQs, in exchange for some other IQs (pertaining to some other functions/skills, where she is NOT an expert).
• But she does not know, who has "created" such IQs!

Can we put in place, some mechanism whereby, all users of Interviewer, can SEE a list of all users who have created questions? — and on which functions/skills?

Perhaps this can be done. When a person downloads any "Candidate Database" from our website:

• We are "harvesting" the questions created by him.
• He is automatically downloading his own/local database of all IQs from our webserver.

I suppose, we have put in place, some mechanism (logic), whereby we are harvesting only "incrementally created" questions, i.e. new/additional questions created SINCE last harvesting.

• So there is no duplication/triplication.
• While harvesting, we know Contributor's name/email ID.

Which question "belongs" to which functional/skill.

Something like:

| Contributor: XYZ | Email ID: --- |

Sl.	Date of Contribution	Function or Skill	Actual Questions
			1. _________
			2. _________
			3. _________

OUR ADMIN TOOL / Harvested Questions shows:

| Year | Count | Cumulative |

| 2008 | | |

Year 2008
Month	Count	Cumu
Jan
Feb
Mar

 

Month: Feb/2008
Dt	Count	Cumu
1
2	30
3	60

So, we are capturing same data from each "contributor" & then aggregating.

What display we now need is:

Interview Question Creators' Peer-to-Peer Network

Dear Recruiter:

If you have created some Interview Questions and if you want to exchange these for some other Interview Questions, created by other users of Interviewer, you can do so here.

Here is a list of all the creators (Function-wise/skill-wise). You can use this platform to approach them with your Exchange Offer, stating:

• What Function/skill questions you want.
• What Function/skill questions you can donate.

To see the list, first select:

• ( ) Non-IT Functions
• ( ) IT Skills

Next select Creators of	[Select Icon]		Next select Creators of	[Select Icon]	No. of Quests
Sales	[x]		Java	[x]
Mkts			Asp
R&D			VB

Here are creators of questions on: SALES

| Sl. | Name | No. of times donated / contributed | Select to send email |

| 1. | Mhatre | 4 | [x] |

| 2. | Patel | 12 | [x] |

| 3. | Venkat | 16 | [x] |

| | Total | (65) | [Open Outlook] |

When OUTLOOK opens,

• Sender's email/name gets automatically entered in [ FROM ]
• Recipients' email IDs get automatically entered in [ BCC ]

Then, sender can compose his message and SEND.

This automatic arrangement ensures that both sender & receivers are authentic / current USER of IndraRecruiter / Interviewer.

We will permit only ONE "function or skill" to be selected at-a-time.

After receiving such a "REQUEST FOR INTERVIEW QUESTIONS EXCHANGE", if one or more recipients respond favorably and actually do send (email) some IQs to the requester thru EXCEL sheets, then the requester would have provision to upload these EXCEL sheets.

In his local database.

Now you may ask,

"Where is the need for this? Every user's, every "created" question, anyway gets uploaded, first, into our CENTRAL webserver database and then next, gets downloaded into each user's local database, as & when he downloads any new PEN.

So, he will get the same set of questions TWICE (or TEN TIMES?),

— once when he requests some other user to send him directly,

and

— next when he downloads PEN (automatic update of harvested questions).

Soon, his harddisk will get flooded with thousands of IQs! Duplicates galore! Very frustrating!"

There is only ONE satisfactory solution. Our webserver must neither harvest any "created" questions nor update user's local databases.

Wrong!

Correct solution is:

• Our webserver will continue to harvest "created" questions, so that we can still generate & display the results shown on p=6 (Exchange Program depends upon this).

But,

• our server will NOT update the IQ databases of users, when they download PEN.

So,

our original 15000 IQ will remain 15000 forever.

Any addition (to his local database) that a user wants to make, will only be thru PEER-TO-PEER EXCHANGE PROGRAM.

Not only will this solution stop duplication, but it will also encourage peer-to-peer sharing, and create a vibrant/dynamic community centered around:

• Interviewer
• IndraRecruiter

If a user (of Interviewer) can keep getting new/fresh questions, contributed by ALL, directly from one/single source, viz: our Webserver, then he has no incentive to "engage" into the "Peer-to-Peer Exchange Program."

By forcing (and facilitating) him to source his questions from SEVERAL SOURCES, we are encouraging the formation of the community. This brings in the next important strategy of Network Externality. We have repeatedly read about the classical examples of phones & faxes. The more there are (on any given network),

the more useful the network becomes. This is also the reason why service providers provide "interconnectivity" between their separate (individual/smaller) networks. In the entire world, there could well be 2000 telephone networks (small & big), but anybody on any network, can easily dial & reach anybody on any other network, in any corner of the earth.

EMI, Virgin, Sony etc. are all (music) record-album companies. I suppose you can log into their websites, select & order "their" own record albums, but not of any other company.

So Apple aggregated many of these in iTune store. Now Apple is planning to do the same with film videos so Apple becomes a central/common depository / Single Source.

But can iPOD users exchange songs amongst themselves? I don't know, but I doubt. Because, if they could, why would they pay Apple (iTune) 99 ¢?

Even if iPOD users CAN exchange songs, do they know, who has got what song?

Turning to our own "Peer-to-Peer IQ Exchange Network", I cannot exchange questions without being a user of Interviewer. Questions can only be created and exchanged thru Interviewer.

So, if I wish to benefit, I must become a user/member, by downloading Interviewer & registering on IndiaRecruiter. Then only do I come to know, who has got what "songs"!

And if I know recruiter-friends who are "experts" in subjects other than my own, then I would encourage them to become members.

Bit-Torrent based networks, also keep track of who is "donating" how much. Then rank members accordingly. Big donors are ranked at the top. Those who only borrow—but rarely donate—are ranked at the bottom.

So, there is a lot of prestige attached to being ranked high. It is also an acknowledgement of your "expert-status" and of your "Social Contribution".

When a donor donates some questions (by emailing), may be we cannot "count" the no. of questions that got emailed. But, I suppose, we can count the "No. of times" that

person/user, donated. If so, we can put a "No. of Times Donated" counter next to each donors name, in the tabulation shown at the bottom of Pg: 6.

May be that tabulation could be even arranged in the "descending" order of that counter.

Although we don't have to attempt to incorporate ALL the features at one go, we need to keep in mind, so that our "Architecture" is open/flexible enough to accommodate these features in future.

Please discuss amongst yourselves what "minimum" can we incorporate right now.

[Signature] 19/01/08

P.S. For consideration

• Someday or other (I hope soon enough!), IndiaRecruiter will become a "Pay-per-click" service for Employers/HR Managers. It is possible that some future service may have to be paid for even by candidates.
• Downloading "candidate Database" by typing PEN in Interviewer, will also become "payable".
• All payments by Corporates, will be in advance, before service-delivery starts.
• Service-delivery will STOP, the moment "Balance in your Account" reaches ZERO.
• "Balance in A/c" can be viewed by the Corpo. Subscriber at any time using his User ID/Password.
• Even if there are 20 users/recruiters/consultants working in a given company, they would like to open only ONE/common account, into which, their ADMIN will deposit payments from time-to-time.

• This would require their ADMIN to share User ID/Password with all those 20 users. Using same User ID/Password, all of them will be able to download "candidate Database (PEN)", using their own /individual/ separately downloaded "Interviewer".
• eg:
• Archana may download Interviewer from IndiaRecruiter.net
• Aparna may download from Download.com
• Sonal may download from Softpedia.com
• etc. etc.
• Rishi may get "Interviewer" tool as an email from Amit.
• Ramya may install "Interviewer" from DIGIT magazine CD
• etc. etc.
• But all of them should be able to download "Candidate Database", using same User ID/Password provided by Thakur.

• Obviously, we cannot ask each of these (one-company) users to:
• Register separately on IndiaRecruiter (as Employers/HR Mgrs)
• Obtain their own separate User ID/Passwords
• Open separate ACCOUNTS by sending separate cheques (or using separate credit-cards online), for (possibly) separate amounts!
• This is simply "unacceptable"—certainly by multi-user companies.
• Downside of this is that all of those 20 users, unknown to each other, may download same PEN (61140?)—and the company may end up paying us Rs. 20/= instead of Re: 1/=.
• But most organisations, may not mind this—especially if each PEN download is as low as Re: 1/=.

However, when it comes to the question of being able to use "Peer-to-peer IQ Exchange Program" is concerned, can we insist that who-so-ever wants to "donate" or "borrow" a question, must be, not only a MEMBER, whose Interviewer has been installed on his local PC and who has, either independently obtained his user ID/Password or has been given User ID/Password by his company ADMIN but such a MEMBER must have also "Created" at least ONE question (on any function-skill).

The idea/philosophy behind this is:

• How can we display your name in the "CREATORS LIST" (see pg: 6), unless you have created (-and uploaded) at least ONE question?
• We don't want "FREE RIDERS", who don't want to create/contribute/share (their knowledge) but, who are only interested in "borrowing"!

Hopefully, such a mandate, will encourage Interviewer-Users to contribute (at least one question), in order to be able to "borrow" from others.

Of course, if I am not interested in "borrowing" any question from anybody, I can still use all the other features of Interviewer. No problem.

Date: 20/01/08

Content:

Now, what happens when Archana / Aparna / Ramya / Sonal etc. all, do “create” some questions & click [ADD QUESTION] button?

Remember, they are all using Thakur’s User ID / Password (ADMIN). On Pg: 6 (Creator’s List), all of these questions, would appear (get listed) against Thakur’s name / email ID!

So, these contributors will not get “recognition”! They will get discouraged. Again, how can they “borrow” from each other? Perhaps, you will come up with some elegant solution.

Date: 22/06/08

To: Rahul $\rightarrow$ Shalaka $\rightarrow$ Shubhangi $\rightarrow$ Poonam

Topic: New Feature - www.ResumeViper.com

Content:

Look-up the “Jobseeker Demo” at your leisure. It is quite interesting what they have done. There is a concept of matching:

• Jobseeker’s “Job Preferences” (i.e., Employer Preferences)
• With Employer’s “Candidate Preferences”

Also see “Resume Distribution Report” (see “who”, “when” & “How Many” cos looked at your resume). Lot to learn from this website.

Date: 23/06/08

To: Rahul $\rightarrow$ Shalaka

Topic: New Feature - Admin Tool for Resume Blast

Content:

I believe Shalaka wants to work on this after uploading. Such an admin tool, should, ideally, give us info. about:

• WHO is sending (Senders)
• to WHOM is sending (Recipients)

Such info would help us to:

• Target the SENDERS (email promo)
• Add (database of) RECIPIENTS (Sachin)

SENDERS Analysis:

Three sides of SENDER-CUBE:

a. IT or Non-IT

b. Designation-Level

c. Current CITY location

Such data will tell us WHO is very active (or very inactive) when it comes to blasting resumes & jobseekers from WHICH region.

Topic: RECIPIENT Analysis

Content:

Three sides of RECIPIENT-CUBE are:

a. Time-axis (Dt./month/year)

b. Country-axis (India – USA – UK – Canada etc.)

c. Type-axis (Corporates or Recru. Agencies)

This analysis will tell us to:

• Add further “Corporates” database for USA
• Add further “Recru. Agency” database for UK etc.

UI Element:

Counter on Resume Blast page, itself should show:

| Category | Value |

| Total Resumes Blasted | [Counter Box] |

| Countrywise Breakup | |

| * India | [Bar/Box] |

| * USA | [Bar/Box] |

| * UK | [Bar/Box] |

| * Canada | [Bar/Box] |

| * Australia | [Bar/Box] |

Note: This data will be a powerful motivator for candidates coming to this page.

Topic: Algorithm Candidate Screening Model Comparison

Source: "Are You Recruiting Well?" - Business Standard / Nov 1, 2007.

Google Algorithm Model	IndiaRecruiter.net (Personal Analytical Recruiting)
Existing employees are assessed on a variety of parameters (team work / experience / academic background etc.)	Software literally mimics human brain as it “reads” each resume, spots keywords, and assigns pre-determined “Weightages” to compute an overall SCORE.
Model determines which traits existing top performers exhibit.	Having computed RAW SCORES, software computes PERCENTILE score and plots a “Frequency Distribution Graph” to display RELATIVE STANDING.
Resumes are screened and given a score from 0 to 100.	Comparison is done vis-a-vis the entire database of all similar candidates.
Candidates called for interview only after initial shortlisting.

 

Topic: Conclusion of Model Comparison

Google Model Objectives:

1. Reduces subjective element in interviews.
2. Increases success rate of choosing right candidates.
3. Saves costs/time for recruitment department.

Note: Google hires innovative people who might be rejected by traditional methods due to academic credentials.

IndiaRecruiter.net Implementation:

Resumes are automatically categorized by score:

• Excellent: $\ge 80$
• Good: $\ge 60, < 80$
• Average: $\ge 40, < 60$
• Poor: $< 40$

Conclusion:

IndiaRecruiter.net achieves all Google objectives and goes beyond by harnessing collective wisdom through a mechanism called POLL-e-Ticks to modify keyword weightages.

Date: 27/04/08 (Page 1/5)

To: Rahul $\rightarrow$ Shalaka $\rightarrow$ Shubhangi

Topic: Resume Search - Functionwise

Content:

When we have processed 6.5 lakh resumes, for each of the 42 functions, we will create a database of TOP 1000 in each function.

Database fields:

• Name of Executive
• Relevance Score
• Function
• E-mail ID

Plan: Develop this search tool in asp.net. If consultants like it, we host it on the website by end of May to generate traffic from jobseekers.

Topic: Resume Search UI Mockup (Page 2/5)

Features:

• Message to HR: "Here you will find text resumes of executives who have not registered on this website... arranged in descending order of Relevance Score."
• Function Selection: (e.g., Accounts, Advertising)
• Results Table: SN, Executive Name, Relevance Score, No. Viewed.
• Action Buttons: Email, Download, Ask Executive to contact me, Edit My Details.

Topic: Email Integration Notes (Page 3/5)

Content:

• Email Action: Clicking "Email" opens Outlook with candidate's ID in 'To' and subscriber's ID in 'From'.
• Messenger Service Template: "Dear [Name], One of our subscribers was searching for executives... with a Relevance Score of [Score], your resume was considered interesting. Our Corporate Subscriber would like you to get in touch..."
• P.S.: Check-out whether you are Underpaid?

Topic: Resume View Alert (Page 4/5)

Content:

Automated email to jobseeker when a corporate subscriber views their resume:

• Subject: Resume View Alert.
• Body: "Your resume was opened/viewed a few minutes ago by [Subscriber Company Name]. Total views so far: [76]."
• Viral Loop: "This is something your friends/colleagues ought to know! Email them the following link: [Hey Guys, you ought to see this]."

• Emphasizes the importance of the Resume View Alert, noting that other portals are "one-way traffic" where seekers simply deposit resumes and hope.
• Predicts this will be a popular feature that may be copied by competitors like Naukri or Monster.
• States that following an alert with a "Contact Me" email would dramatically boost reputation among jobseekers.

• Draft for a "Laughter Club" section on the website.
• The author, Hemen Parekh, mentions cartoonists like Mike Flanagan (HAMBONE), Salam (BLINKERS OFF), and S. Adams (DILBERT).
• The section is intended for visitors to enjoy and share with colleagues or bosses.

Date: 23/08/08

To: Rahul, Poonam, Sarang, Shalaka, Shubhangi, Diya

• A graph illustrating the relationship between "Ave. Time on Site (Min)" and "Pages/Visit (no)."
• The note states, "This relationship is fairly obvious."

• Identifies the primary problem as candidates not knowing about IndiaRecruiter or being unmotivated to submit resumes.
• Table for popularizing the website:
• Corporates to Recm. Agencies via Vacancy Blast.
• Ourselves to Candidates via Resume Blast and email campaigns.
• Corporates to Candidates via Job Alerts.

• Proposal for job alerts: as soon as a corporate posts a job advertisement, a message is forwarded to matching candidates.
• The message will appear to come from the Corporate Subscriber rather than IndiaRecruiter.net to increase effectiveness.

• Workflow for Job Alerts:
1. Corporate posts a job.
2. IndiaRecruiter displays matching resumes (Match-Index).
3. Corporate selects/reads resumes and clicks "Send Job Alert."
4. Message is emailed to the candidate.

• Defines the Match Index formula:

$$Match Index = \frac{\text{"Score" based on keywords found in each resume}}{\text{"Score" based on keywords found in the Job-Description of that job-advt}}$$

• Identifies a potential "disaster" where short job descriptions with only one keyword result in skewed Match Index percentages like $600\%$, $800\%$, or $1500\%$.

• Expresses concern that unrealistic Match Index scores will cause the service to lose credibility with Corporate Subscribers.
• States subscribers will only believe match requirements in the $60\%$ to $80\%$ range.
• Suggests this can be achieved if the number of keywords in the Job Description is greater than the number of keywords in any given resume for a specific skill.

Subject: Match Index Logic

This can be done as follows:

• Ignore keywords contained in the job description.
• As soon as a Subscriber selects any particular "Function" (while posting a job-advt), populate/fill-up the "skills/knowledge (keywords)" box with ALL the MASTER keywords in our Database, pertaining to that function/skill.
• Then compute "Match Index" as:

$$\text{Match Index} = \frac{\text{Keywords (score) found in Resume}}{\text{Keywords (score = 100) in MASTER set}}$$

• This will always be less than 100!

Subject: Keyword Harvesting

Of course, we should allow the Subscriber to "Add/Delete" the filled-in Keyword Box.

• Against each FUNCTION / SKILL, we will store (in a separate table) the "keywords" finally retained in the keyword box by Subscribers.
• Not only will we be able to "harvest" totally NEW keywords (not found in our MASTER-SET), but we will also be able to compute the "Frequency of Usage" for each keyword and arrange them in descending order for the next time they are displayed or filled up.

Subject: User Experience Flow - The "Twist"

Next twist:

We should not immediately display the "Matched Resumes" (with their Match-Indices). We should display this tabulation only AFTER the Subscriber clicks (at the bottom of the Job Ad Posting):

[Send Job Alert To matching Candidates]

If he clicks on this button, then we display the following:

"Your Job Advt has been emailed to the following 'Matching' Candidates"

Sl No.	Match Index (Descending order)	Name	...	...

 

Subject: Verification and Backend Sorting

Such a display will "Assure" the Subscriber that we DO have suitable resumes (and that we are not simply bluffing!). He can even click open those resumes and satisfy himself. This could be VERY VERY assuring if those very keywords get highlighted (red) when the Subscriber opens any resume.

Of course, all of this becomes possible only after "Resume Sorter" sorts 700,000 resumes and creates function/skill-wise MASTER folders, containing 700 top-rated resumes. On-the-fly, our s/w can possibly "Match" only 700 resumes against any job-advt.

(Signed/Dated: 06/04/08)

Subject: Revising "Submit Resume" Form (Dated: 30/07/08)

To consider for incorporation:

1. Currency Converter
2. SKYPE ID (optional)
3. Any field in PERSONAL INFO block that should be made mandatory? Quite often, in "Profiles", we see the PERSONAL INFO block nearly empty! Looks bad.
4. For a jobseeker, option to select:
• Make my resume accessible in Resume Search.
• Keep my resume PRIVATE [PRIVACY REASONS] (clicking on this link will open "PRIVATE RESUME" page).

If a jobseeker selects this option, we may have to devise/allot a PEN (Permanent Executive Number) which is distinctly different, so that such PEN do not show up in any kind of Resume Search. A simple "MISSING" Number from existing PEN series may not suffice.

Subject: Private Resume Features (Dated: 30/07/08)

Dear Executive:

If you are a Senior Executive, you don't want your online resume to be visible to anyone and everyone—anytime. On a case-to-case basis, you want to decide:

• Who should see it
• When
• For what purpose

Example: A headhunter approaches you with an interesting vacancy and requests your resume:

• Give him your REN (Permanent Executive Number) over the phone. Ask him to enter this number in our Interactive Interview Tool and download.
• Locate his Company-Name in our Resume Blast, tick against that name, and click [Send Profile]. Or email it to him yourself.

Subject: Privacy and Remote Interviewing

A Senior HR executive of a respected company wants to interview you. Obviously, you do not wish to meet him in his office—or even in a hotel, where you can be spotted!

No problem:

Give him your PEN over the phone. When he enters that in the Interactive Interview Tool, he will see your SKYPE ID and conduct a Video-Interview on your home PC. Your privacy is guaranteed!

Your Advantages:

• When you choose the "PRIVATE RESUME" option, none of our Corporate Subscribers can search/locate/read your resume through online Resume Search. It is simply invisible to them.
• You can periodically update your resume through "Edit Resume."

Subject: The Long Tail Keyword Search (Mockup)

Message to HR Manager:

"The Long Tail" is a fascinating book by Chris Anderson. You may be looking for candidates possessing some UNIQUE skills (Rare Gems). How to locate those resumes amongst millions in a few minutes? Simple. Keep going down the Long Tails or just type in. — Hemen Parekh

Interface Elements:

• Keyword Occurrence Frequency Table: Ranges from 10 million down to 0-10.
• Search Box: "For that 'Rare Gem' (Unique skill denoting keyword) type here [______]".
• Results Table: Includes columns for Executive Name, Designation, Company, Number of times resume was viewed, and a link to open the resume.