WEBVTT

00:00:02.400 --> 00:00:06.300
I first worked on this kind of problem
during a technical assessment for a data

00:00:06.333 --> 00:00:10.333
analyst position at an international
educational publishing company.

00:00:10.800 --> 00:00:15.300
The task was to examine a product catalogue,
make sense of its pricing structure,

00:00:15.333 --> 00:00:18.200
and explain which findings were worth pursuing.

00:00:18.800 --> 00:00:23.600
I later rebuilt the exercise as an
independent case study, with fictional titles,

00:00:23.633 --> 00:00:28.933
series and prices, so I could share
the analytical process publicly.

00:00:30.300 --> 00:00:33.300
A price list looks simple
until you look closely.

00:00:33.767 --> 00:00:38.033
A single course might have a student book,
a workbook, an answer-key version,

00:00:38.067 --> 00:00:41.733
a digital licence, several
levels and multiple editions.

00:00:42.200 --> 00:00:45.133
Two similar names can
describe different products.

00:00:45.300 --> 00:00:48.867
If you compare their prices without
accounting for those differences,

00:00:48.867 --> 00:00:53.367
an apparent anomaly may tell you more
about the catalogue than about the pricing.

00:00:53.833 --> 00:00:56.733
So the first question was
not which product costs more,

00:00:56.767 --> 00:01:00.367
but which products can fairly be compared.

00:01:01.700 --> 00:01:08.600
For this public version, I generated 362 fictional
product records across 12 fictional series.

00:01:08.767 --> 00:01:12.300
I took five raw fields and enriched
them into nineteen attributes,

00:01:12.333 --> 00:01:17.267
covering product type, level, edition,
bundle content and digital access.

00:01:17.733 --> 00:01:22.067
I preserved the source and kept every
correction in a separate reporting view.

00:01:22.300 --> 00:01:28.600
Then I grouped products by their actual
configuration, so that only the level could change.

00:01:29.933 --> 00:01:34.200
That gave me 60 comparable product
groups spanning more than one level.

00:01:34.667 --> 00:01:37.200
54 had the same price throughout.

00:01:37.233 --> 00:01:40.933
Six did not. Four involved
a cheaper Starter level.

00:01:41.400 --> 00:01:44.767
That could reflect an
intentional entry-level policy,

00:01:44.800 --> 00:01:47.833
but a list price cannot prove why it exists.

00:01:48.300 --> 00:01:53.000
The remaining two differences called
for a different kind of investigation.

00:01:54.333 --> 00:01:56.833
The clue was inside the groups themselves:

00:01:56.867 --> 00:01:59.467
each one listed the same level twice.

00:01:59.633 --> 00:02:04.800
In both cases, the product title named one
series while the series field named another.

00:02:05.267 --> 00:02:08.333
Grouped by that field,
it looked like a price difference.

00:02:08.667 --> 00:02:13.800
Checked against the product's identity,
it was a filing mismatch, not a pricing decision.

00:02:14.267 --> 00:02:19.367
I reassigned those two records in the reporting
view, leaving the original data untouched.

00:02:19.833 --> 00:02:27.000
With the same 60 comparable groups, the number
with consistent prices rose from 54 to 56.

00:02:28.333 --> 00:02:30.733
The catalogue had other traps too:

00:02:30.833 --> 00:02:33.033
two spellings of the same edition,

00:02:33.067 --> 00:02:36.367
a title that contains the
letters e-book but is printed,

00:02:36.400 --> 00:02:40.367
an unusually high price attached
to an ambiguous product code,

00:02:40.400 --> 00:02:44.500
and books sold in two parts that
cost more than the complete volume.

00:02:44.967 --> 00:02:49.533
They were prompts for investigation,
not excuses to invent an explanation.

00:02:50.000 --> 00:02:54.000
There are no sales, cost,
margin or demand data here,

00:02:54.033 --> 00:02:58.367
so the analysis does not recommend
a price or claim a financial impact.

00:02:58.633 --> 00:03:02.300
Every figure in this demonstration is fictional.

00:03:03.633 --> 00:03:04.700
The lesson is simple:

00:03:04.733 --> 00:03:09.367
before making a decision about a price,
make sure you know what product you are looking at.

00:03:09.833 --> 00:03:14.067
You can explore the dataset, inspect the
calculations and download the case study

00:03:14.100 --> 00:03:15.467
materials on the website.

00:03:15.500 --> 00:03:19.000
I'm Alfonso Cifuentes, and this is Pricing Demo.