Licensed to Muhammad Asim, Lahore University of Management Sciences ([email protected]). Do not copy or distribute.

CASE STUDY

ABCDEMAIL:
SEARCH ENGINE ANALYTICS
ARVIND RANGASWAMY AND AMIR ALI KANI

Copyright © 2020 by DecisionPro Inc.

This document is primarily intended to be used in conjunction with the Enginius

software suite. To order copies or request permission to reproduce materials, go to
http://www.enginius.biz. No part of this publication may be reproduced, stored in a
retrieval system, used in a spreadsheet, or transmitted in any form or by any means
–electronic, mechanical, photocopying, recording or otherwise– without the
permission of DecisionPro, Inc. v200612

1
Licensed to Muhammad Asim, Lahore University of Management Sciences ([email protected]). Do not copy or distribute.

Background
ABCDEmail (a disguised email marketing company) is a major provider of software for email marketing. The
company’s software and services enable firms to manage their email campaigns on an ongoing basis. Using
the software, companies can create eye-catching email campaigns with pre-designed templates (regularly
updated with fresh designs) or by developing their own templates. The software also enables companies
to automate many email management and delivery tasks, such as generating and maintaining contact lists,
sending automated welcome messages to new customers, thank you email to current customers, or
greetings on birthdays and other occasions, win-back emails, and A/B testing of alternative email subject
lines or other features. The analytics associated with the software allows users to track and evaluate in real
time the effectiveness of responses to email campaigns. The software also provides integration tools to link
email marketing with CRM systems, shopping carts, social media campaigns, surveys, etc.

ABCDEmail uses Google paid advertising for customer acquisition. The company had spent over $650,000
from October 30, 2015 to October 29, 2016 on search advertising campaigns. On November 25, 2016,
Jerrod Smith, Advertising Manager for the company, was working on preparing a report summarizing the
highlights from the search ad campaigns for the previous year and outlining the search engine strategy for
the following year. An analysis of past data suggested to Jerrod that the lifetime value of a typical customer
acquired via search averaged around $300 in paid subscription revenue net of costs of servicing the
customer (lifetime was calculated over a 3-year period).

He was scheduled to meet with the CMO of the company, Rita Hewson, on December 6th. At a previous
meeting, Rita had expressed some concerns about whether the search ad campaigns were costing more
than the returns the company was generating and had posed the following questions to Jerrod.

Assume the role of Jerrod Smith, and put together a report for Rita Hewson.

This case requires downloading the data set ABCDEMAIL DATA and using
Microsoft Excel for the needed analysis. When you open the ABCDEmail
 case on Enginius, you will see the needed Excel file listed in the Quick
Links section.

Case study questions


Question 1
Is it worthwhile for the company to continue using branded keywords (these are
keywords in which “ABCDEmail” is included in the keyword)? [Use spreadsheet DB1
to assess the costs and benefits (e.g., traffic and profits) generated by branded and
unbranded keywords].

2
Licensed to Muhammad Asim, Lahore University of Management Sciences ([email protected]). Do not copy or distribute.

Question 2
Was the company using the right mix of broad match, phrase match, and exact
match keywords? It is sufficient to provide directional guidelines here, rather than
precise recommendations. [Use spreadsheet DB1 to explore this issue].

Question 3
What, if any, are the financial benefits to be gained by dayparting (i.e., choosing
specific hour(s) of the day for the search campaign)? Were there any benefits to
concentrating the ad spend during the weekdays when business managers are
likely to be working in their offices? [Use spreadsheet DB2 to explore this issue].

(Advanced optional question: Answer this part only if your instructor asks you to
undertake an analysis to assess keyword performance by various dayparts such as
weekday workhours, weekday evening hours, weekday nights, weekends, etc. Use
INDEX and MATCH functions in Excel along with PivotTable command to convert
hourly data into data corresponding to various dayparts.)

Question 4
Rita had just returned from a conference where several firms reported they were
getting search term efficiencies around 50%, although there seems to be a wide
range from about 35% to 75%. The “search term efficiency” of a paid search
advertising campaign is the “cost of search queries (i.e., search terms) that were
clicked and converted at least once, divided by the total cost incurred for all search
queries that triggered a keyword match”. Based on this definition, a search term
helps improve efficiency if it converts at least once during the entire campaign.

What is the “search term efficiency” index for ABCDEmail? [Uses spreadsheet DB3
to explore this issue].

How would you modify the set of keywords used in the campaign to increase
search term efficiency (Be specific)?

Question 5
What should ABCDEmail do to be smarter about the keywords or types of
keywords to bid on so that it will at least break-even in terms of the lifetime value
of acquired customers? Develop a set of operational and strategic guidelines to
help the company select keywords for its future campaigns.

Question 6
Overall, is our search engine advertising budget adequate? (i.e., are we over-
spending or under spending?) [Use spreadsheet DB1or DB3 to explore this issue].

3
Licensed to Muhammad Asim, Lahore University of Management Sciences ([email protected]). Do not copy or distribute.

Notes

(1) A conversion does not necessarily mean a subscription to the service. But without a conversion, there
is no subscription that could be attributed to search advertising. Based on his experience, Jerrod believed
that anywhere from 2 to 4 conversions will translate into a paid subscription, and not just a free trial
subscription.

(2) Before computing search term efficiency, first explore the data in spreadsheet DB3. The “SearchQuery”
column refers to the actual query (search term) used by a potential customer. For example, in the first row,
the search term typed by the customer was, “free email templates,” whereas Google matched that query to
the keyword, “html newsletter templates” using the “broad” match criterion. When search term efficiency is
high, it means that the keywords used in the ad program helped convert a greater proportion of the search
terms that were associated with clicks or visits (i.e., at least one conversion) via the paid search campaign.
If you can get rid of all the non-performing search queries by selecting an appropriate set of keywords, you
will get an efficiency index of 1 (an unreachable goal in most instances).

Within AdWords data, the impressions associated with search terms are a sub-set of the impressions
associated with keywords. For keywords, impressions refer to the number of times an ad is served for that
keyword, whereas for search terms, impressions are counted only if the search term receives a click. So,
search terms that did not generate any clicks do not contribute to aggregate impressions even though an
ad might have been served with that search term. As a result, you will see some differences in cumulative
statistics (e.g., sum of impressions, sum of costs) computed from the three different spreadsheets for the
same keyword or set of keywords].

Acknowledgement: We thank Scott Ellis and Alan Hamor of AdWorthy for the data used in this case. The case does
not intend to serve as endorsement, source of primary data, or illustration of effective or ineffective handling of a
managerial situation. Certain data and names have been disguised to protect the strategic interests of the
participating company.

4
Licensed to Muhammad Asim, Lahore University of Management Sciences ([email protected]). Do not copy or distribute.

Appendix

Description of data fields in each spreadsheet

Tutorial for using Excel Pivot Tables to extract data from the Excel spreadsheet
(ABCDEmailData(Enginius.xlsx)).

A pivot table enables you to extract relevant data from a large data set by grouping data into different
specified categories. To generate pivot tables, select the data of interest in Excel, then select Insert on the
main menu, and then select PivotTable. In selecting large data sets, you can use the Cntrl-Shift-Home or
Cntrl-Shift-End keys. You can also use Cntrl-Shift-Arrow keys to select specific columns or rows of numbers.
The following screen shots show what you see if you do these steps correctly.

5
Licensed to Muhammad Asim, Lahore University of Management Sciences ([email protected]). Do not copy or distribute.

Click OK on the Create PivotTable dialog box. You can drag and drop fields into the four quadrants of the
PivotTable Fields as shown below. Use the down arrows in the  Values quadrant and click on Value Field
Settings to select the type of computation you would like for each of the variables listed in that quadrant.
Here, we chosen to compute the Sum of Impressions, Clicks, Conversions, and Cost, and the Average value
for the variable AvgPosition. Note also we use “MatchType” as an additional Filter for displaying the data.
This means you can further sub-select only “Broad”, “Exact” or “Phrase” matched keywords for displaying the
computed values, or for further analyses.

6
Licensed to Muhammad Asim, Lahore University of Management Sciences ([email protected]). Do not copy or distribute.

You will then get a tabular output (a part of which is shown below).

You can use Excel manipulations on this pivot table, for example, to select branded versus unbranded
keywords for further analysis. You can also use Excel functions to compute other metrics such as Cost per
Thousand Impressions (CPM), Click through Rate (CTR), Cost per Click (CPC), or Cost per Action or Conversion
(CPA) for each keyword.

The following PivotTable Field settings will generate data organized by the three different “Match Type.”

The Date field in each database is important for several additional types of analyses, particularly for doing
panel data analysis (a special type of regression model) that is useful when we have multiple observations
for the same entities (e.g., keywords) on several occasions (e.g., different time periods). You can extract the
day of the week corresponding to a date using the Excel command =TEXT(cell, “dddd”), and the month
corresponding to a date using the Excel command =TEXT(cell, “mmmm”). The following example illustrates
how you do this:

7
 Licensed to Muhammad Asim, Lahore University of Management Sciences ([email protected]). Do not copy or distribute.

An example panel data derived from spreadsheet DB1 is included in the Excel file
ABCDEmailData(Enginius).xlsx, and is also distributed online at Enginius.biz along with the case. This data
was generated using the following Pivot Table Fields after each Date value was replaced by the
corresponding month, and after some rearranging of the resulting pivot table.

In this panel data, the performance of each keyword is aggregated by month, and a deeper exploration of
this data using panel data analytics is useful for developing strategic guidelines for ABCDEmail. Panel data
analytics with this extracted data can provide you insights about the systematic variations in the intrinsic
value of each keyword (even among a set of seemingly profitable keywords), after accounting for other
factors that could also influence performance (e.g., match type, the position at which the impressions for
that keyword occurred, etc.). Other types of panel data could be obtained from the three databases
provided to you. Some examples are: (1) MatchType as panel entities and HourOfDay providing the multiple
replications, or (2) HourOfDay as the panel entities and MatchType providing up to three replications for
each hour of day

Powered by TCPDF (www.tcpdf.org)