WEBVTT

0:00:02.900 --> 0:00:08.790
okay so I'm going to take you through a
short presentation of the Erlang

0:00:08.790 --> 0:00:15.420
calculator and first of all, I just want to
explain what an Erlang calculator is. In

0:00:15.420 --> 0:00:20.460
essence it's very simple. It works out the
number of people that you need to handle

0:00:20.460 --> 0:00:26.550
a given number of phone calls. So for
instance you've got 200 phone calls an

0:00:26.550 --> 0:00:32.489
hour. It will work out how many people you need in a call centre to handle that number

0:00:32.489 --> 0:00:38.010
off calls. So it's quite straightforward
quite a straightforward concept and very

0:00:38.010 --> 0:00:45.090
useful thing to use if you're planning
in a contact centre. So why is it called

0:00:45.090 --> 0:00:48.989
Erlang? It's called Erlang as its named after

0:00:48.989 --> 0:00:55.200
the Danish mathematician AK Erlang who
developed the formula over 100 years ago

0:00:55.200 --> 0:01:01.350
in 1917, and here's a copy of his
original paper which was really the

0:01:01.350 --> 0:01:07.049
landmark paper and it's amazing that's
still so fresh and vibrant over 100

0:01:07.049 --> 0:01:11.579
years on and basically works at the
probability that call has to wait in the

0:01:11.579 --> 0:01:16.139
contact centre and it can convert the
number of calls, something called the

0:01:16.139 --> 0:01:20.310
average handling time, to go into a
little bit better and a service level

0:01:20.310 --> 0:01:26.880
into the number of staff that you need.
It is very robust and it is in

0:01:26.880 --> 0:01:32.630
very widespread use. It's based on
something called the Erlang C formula.

0:01:32.630 --> 0:01:39.280
Now the Erlang C formula, here's the maths involved behind, it fact it's part of the maths,

0:01:39.300 --> 0:01:46.319
believe me the maths does get very scary,
very quickly, so it can get quite complex

0:01:46.319 --> 0:01:52.109
you need to be able to do series,
factorial numbers, exponents and so on.

0:01:52.109 --> 0:01:56.039
But luckily for you you don't have to
get into all of that maths because an

0:01:56.039 --> 0:02:01.109
Erlang calculator does all of that for
you and basically what the Erlang

0:02:01.109 --> 0:02:06.779
calculator does is it takes a number of
inputs. So things like the number of

0:02:06.779 --> 0:02:12.290
calls per period, the Average Handle, the
average time to handle,

0:02:12.290 --> 0:02:18.230
which was known as the average handling time and a service level target and then

0:02:18.230 --> 0:02:23.060
based on that it works out the number of
advisers that is needed. So in terms of

0:02:23.060 --> 0:02:27.410
the inputs typically what you do is you
can work out the number of calls

0:02:27.410 --> 0:02:32.510
throughout a period in this case we've
set it here to 30 minutes. You work out

0:02:32.510 --> 0:02:37.340
what the average handling time is, now
that effectively is a combination of

0:02:37.340 --> 0:02:43.790
both the average time it takes to handle
a telephone call but also or the average

0:02:43.790 --> 0:02:48.740
amount of talk time, but also the average
amount of paperwork that is required

0:02:48.740 --> 0:02:53.060
after that call that is specific to a
call. You put the two together

0:02:53.060 --> 0:02:58.310
and it's called the average handling
time and then you put in a service level

0:02:58.310 --> 0:03:05.120
target, so for instance answering 80% of
calls in 20 seconds now 80% of calls in

0:03:05.120 --> 0:03:10.010
20 seconds is the is the typical
standard figure if you like use within

0:03:10.010 --> 0:03:14.630
the contact center industry a lot of the
better contact centers of late have been

0:03:14.630 --> 0:03:21.590
adopting a model of answering 90% of
calls within 15 seconds. Now some people

0:03:21.590 --> 0:03:26.330
say well why can't we answer 100% of
calls in 4 rings,  that's typically

0:03:26.330 --> 0:03:30.260
something that a chief executive will
come down, so we're gonna answer 100% of

0:03:30.260 --> 0:03:37.220
calls within 4 rings well the problem
with that is that probabilities don't do

0:03:37.220 --> 0:03:43.280
100%, you can put in 99%. But if you're
gonna have 99% of calls answered in 4

0:03:43.280 --> 0:03:47.360
rings you need north lot of awful lot of
people because sometimes you get calls

0:03:47.360 --> 0:03:52.790
bunching up and a lot of the time it can
be a bit quiet so a good rule of thumb

0:03:52.790 --> 0:03:57.470
is 90% of calls in 15 seconds or a 80% of calls in 20 seconds you can

0:03:57.470 --> 0:04:02.180
adjust that and then you can work out
the number of advisers needed. That's all

0:04:02.180 --> 0:04:06.200
built into the and I calculate we'll do
a demonstration in short while and to

0:04:06.200 --> 0:04:10.250
give you a rough idea of some of the
outputs that come through here some of

0:04:10.250 --> 0:04:15.140
the here are some of the outputs work
out the number of agents what the

0:04:15.140 --> 0:04:19.880
percentage they're likely to be answered
within a certain threshold something

0:04:19.880 --> 0:04:24.620
called the occupancy. We don't need to
get into that quite yet and also the

0:04:24.620 --> 0:04:27.500
number of calls
and then it produces a graph showing the

0:04:27.500 --> 0:04:31.790
number of people you needed and also
what happens if you adjusted the number

0:04:31.790 --> 0:04:37.190
of people up and down what would happen to the service level it's a very

0:04:37.190 --> 0:04:43.910
straightforward nice easy answer. One
extra level of complication is that

0:04:43.910 --> 0:04:49.010
because the number of calls vary across
the day it also has a day planner

0:04:49.010 --> 0:04:54.860
built-in that takes account of the
variation of contacts across the day. So

0:04:54.860 --> 0:04:59.120
for instance in lots of call centers it
gets busy mid-morning gets also busier

0:04:59.120 --> 0:05:03.470
during the afternoon slackens off a bit
lunch time and then it tails off across

0:05:03.470 --> 0:05:08.600
the course of the day. Now that does 
very much depend from one contact center

0:05:08.600 --> 0:05:16.000
to another so that is a quick
introduction to an Erlang calculator

