MATCH INDEX & YIELD RATE

Abhi → Rajeev

23-02-05

1/6

Match Index & Yield Rate

▸ Nimit is meeting Hemant Sachdev (Airtel) today. He may also meet Manoj Dawne. Airtel are slowly moving towards a nation-wide roll-out of JAM. In a large organization like Bharti, things always move slowly since many depts/individuals are involved. Our job is to keep up a gentle pressure all the time, without offending/irritating the individuals concerned. These comments apply to all MSPs — and more so to Govt. org. like MTNL/BSNL.

▸ Our priorities are as follows: ➔ In quick succession, tie-up with all MSPs

➔ Add as many Indian/American job sites as possible

➔ Keep increasing the QUANTITY of job-ads downloaded everyday

➔ Keep improving the QUALITY of match-making, thereby increasing the RELEVANCE of alerts.

▸ This is where, daily/weekly/monthly compilation of MATCH-INDEX comes into picture. It is an indicator of RELEVANCE.

 

What is Match-Index?

Let us say, a jobseeker has asked us to deliver to him 3 alerts every day i.e. 30 x 3 = 90 alerts/month.

Now, during the course of the month, we succeed in delivering to him, a total of 60 alerts only.

Then, Match-Index = 60 / 90 = 66% (say for March 2005)

But, in April 2005, we succeed in delivering 70 alerts.

Then match Index for April 2005 = 70/90=77%

In case Mr. Mahatre, edits No. of job alerts to be delivered by per day at anytime software will automatically adjust the denominator.

Now, we can fix some rules artistry viz:

 

A match-index of less than 66% is UNSATISFACTORY

▸ A match-index between 60% to 80% is GOOD

▸ A match-index above 80% is EXCELLENT

When, for each user, a monthly graph as shown above gets plotted (— of course, automatically without any human intervention), then we can clearly see whether, for that person:

• Match index shows steady IMPROVEMENT

• " " STAGNANCY

• " " DETERIORATION / DECLINE

If Match Index shows DECLINE, then that case needs immediate investigation (WHY?)

We can even club-together, all cases, which, during a month, showed a decline investigate as a group

Could this decline be due to –

Very stringent / difficult job performance criteria set by those candidates?

OR

There is significant drop in our daily downloading of job advt ?

 

OR

• Both

If such a "decline" is observed in case of a very few candidates, we should find time to phone them & inquire personally.

If such a "Decline-Group" consists of hundreds of candidates, we can send out to them, some automatic-event-triggered email, saying, "This is what we have observed/noticed in their case. Is there anything that they would like to suggest to reverse this trend?"

Some similar automatic email can also be delivered to any candidate, in whose case, Match-Index remains STAGNANT (at same level) for more than 3 consecutive months.

Of course, in those cases, where there is, • Constant/continuous improvement

OR

• Statistically "random" fluctuation (up & down),

We need not send any email.

I feel, data complication ( People wise) & graph plotting can be completely Automated. Initially if we so desire we may not plot a graph but simply plot a table.

Mr. A. J. Mhatre		PEN:
Month	No of alerts desired	No of alerts delivered	Match Index = b/a
	a	b

 

Initially, to further simplify/reduce workload, we can simply email such a tabulation to ALL jobseekers, irrespective of whether they are belonging to Decline/Stagnant or Improvement Groups. This email can get triggered on 3rd of each month – of course, with some appropriate message/covering letter (which I can draft & give you).

We can bring-in Group-Wise refinements later on.

But receipt of a monthly tabulation (month-after-month), in itself, will be a pleasantly surprising & a happy/different "EXPERIENCE" for all jobseekers!

In one more way, we will be telling them,

"We are NOT a typical jobsite!"

Remember, our JAM is a "Service" and a service can only be "experienced" – it cannot be touched or smelled or be heard , like in a case of product!

We have to constantly find, ways means to differentiate ourselves from all jobsites.

It is as Jack Trout says,

Differentiate or Die!

As far as YIELD-RATE is concerned, it is purely for our own benefit/utility/performance-monitoring.

Yield Rate = (No. of "Apply Online" clicks in a month / No. of job-alerts delivered in a month) × 100%

Yield-rates should be calculated for:

• Each individual subscriber (Individual Yield)
• For all subscribers belonging to a given MSP (MSP-wise Yield)
• For all subscribers belonging to all MSPs put together (Gross/Aggregate monthly- or daily-yield)

Of course, all data must be captured DAILY. Then we can manipulate data in many different WAYS & for many different periods and for any combination of MSPs or Circles or Geographic regions etc., etc.

Yield-rate is the bottom-line which tells us:

"How much money are we making?"

• daily/weekly/monthly/quarterly/annually?
• from MSP A or B or C or combination?

23/02/05