Formula is right - but result looks wrong

[Sir Babydum GBE]

Hi

Can someone help explain the attached to me please?

I have to provide stats regarding two people. Agents 1 and 2 take calls in a call centre - and try to succesfully answer enquiries without transferring the call.

The number of calls is recorded as is the number of transfers.

My problem is that I get a percentage of calls transferred figure vor each agent each week. and we calculate a percentage point increase/decrease over the previous week. The combined percentage point increase for both agents is greater than the highest individual agent increase.

How can this be possible? People will question me when I send the details out - I don't know how to answer.

[rory]

They are not percentages of the same thing so the percentages are not really related.

[Sir Babydum GBE]

They are not percentages of the same thing so the percentages are not really related.

Well, they're all percentages of call vs transfers - it's just that we have combined calls vs combined transfers for the grand total - so I'm still not sure how the combined % increase can be greater than the highest individual % increase.

Am I being thick?

[Bob Phillips]

The percentage values depend upon the volumes done by each agent. So as agent 1 takes far more calls than agent 2 as well as answering more of them you are getting a weighted figure, and not excluding the weighting in the overall results.

What you need to do is analyse the numbers from a common base, say 100 calls received.

From this you would see that the analyse would look like

100 Weighted calls

Agent 1 Calls Answered 35.59 39.81
Agent 2 Calls Answered 56.21 53.69


From this the calculated percentages would be

Agent 1 % Answered % 35.6% 39.8% 4.2%
Agent 2 % Answered % 56.2% 53.7% -2.5%


Note that this is the numbers that you got, so we are correct so far.

Then the summary per 100 calls per received is

Total Calls Received 200 200
Total Calls Answered 91.81 93.50

which gives you a comparative analsyis of

Answered v. received 45.9% 46.8% 0.8%

[rory]

Agent 2's success rate is down in week 2 but he is still more successful than agent 1 and in week2 he is handling more calls, so the overall success rate is up.
For each variance you are subtracting a percentage of one amount from a percentage of a different amount, to get a percentage point number, which isn't really a percentage of anything. You are then trying to compare the three variances even though they don't actually have anything in common.
Imagine you had 10 calls, 5 to each agent, and agent 1 is 0% successful and agent 2 is 100% successful. Total successes = 5, i.e. 50%. Next week all calls are handled by agent 2 who is 90% successful. Total successes = 9, i.e. 90%. BUT agent 1's percentage point change is 0, and agent 2's is -10%, and the overall is +40%. The numbers are not related to each other.
Does that help at all?

[Bob Phillips]

I have knocked up an example, with a simple dashboard.

You have to understand how the figures are calculated if you are going to present them, so you need to understand the basic maths/statistics behoind them.

[Sir Babydum GBE]

I have knocked up an example, with a simple dashboard.

You have to understand how the figures are calculated if you are going to present them, so you need to understand the basic maths/statistics behoind them.Don't tell anyone Bob, but i think I'm thicker than I thought I was.

But that helped a lot.

Cheers

Sir BD

[Bob Phillips]

I think it is one of those things that is obvious ... once someone has explained it to you. Until then, you think you can compare them, after all apples and oranges are all fruit, right?

I know that most management are incredibly stupid, like the government they want simple, once sentence answers to everything, and are incapable of understanding the true complexities (or even the true simplicities) behind everything, which is why it is so easy to fool them with a few biased numbers.

[Sir Babydum GBE]

Thanks again Bob

And thanks Rory - somehow I missed your answer until just now... It makes sense though.

I may use it, or I may tactfully say to my managers something like "I'd like to explain it to you, but I think you're probably a bit too stupid to understand - so if you just trust that my figures are right, the office will be a nicer, less tense place to work. PS, where's my pay raise?"

What do you reckon?

[Bob Phillips]

Sounds like a well constructed, perfectly objective, irrefutable argument to me.

But seriously, I am sure that you are experienced and savvy enough to know that you can say those things without saying them. One liners will go down well if the message is one they want to hear, so if you are happy imparting the one liner that 'things are getting better' then go for it. If you are cnveying less positive news, you know that you will be challenged, so you have to have backup facts, and an understanding of the metodology, to withstand the onslaught (most managerment that I have met think that aggressiveness is strong management - idiots!). You also have to have those extra facts and understanding in case someone asks questions EVEN when you are conveying a positive message. Moral, you have to understand the methodology, and don't tell the whole story at the start, save something for the almost inevitable questions. Silly I know, but work is political, just like life.