Professor : Keyoor Purani

PREPARED AND SUBMITTED BY

From the case, the optimal search marketing campaign is based on choosing effective
allocation of ad dollars across the various search engines, as well as selecting appropriate
keywords and bid strategies for placement on the search result page for Internet users. In
determining the optimal strategy, the case presents background information on the airline
industry as well as the Internet search options available at the time, including Google,
Microsoft MSN, Yahoo!, and Kayak. Additionally, background information is provided on
SEM and its associated costs and means of measuring the successfulness of each marketing
effort. The table Exhibit 6 included line item information for keywords and keyword
groups for each campaign and search engine, implying that strategies between each search
engine is at least slightly difference, although high level conceptually they all lie on similar
principles .

However in another case, the author suggests that buyers using search engines to look for
information tend to trust and follow links displayed in the editorial section of the search-
results page. Most on-line sellers, however, do not invest in search engine optimization
(SEO) to get higher search-results rankings for their listings, but instead prefer paid
placements. They explain that SEO is more expensive than paid placements, produces
results that do not justify its cost, and does not consistently lead to high search-results
rankings. This implies that sellers would invest in SEO if it were less expensive and its
rankings were more consistent. However, even if SEO and paid placement cost the same
and SEO always produced high rankings, paid placement would still be the search engine
marketing (SEM) strategy of choice for most on-line sellers. Interestingly, no optimal SEM
strategy includes SEO. These findings are especially significant for advertising
professionals who have to justify investing in paid placements despite overwhelming
evidence that buyers ignore them and follow links in the editorial section of the search-
results page. In the PivotTable we can isolate and separate variables that lead to deeper
insight into the impact and revenue effect different publishers have on facilitating Air
France ticket sales.

Placing Publisher name in ‘Column Labels’, with ‘Campaign’ and ‘Keyword’ as row labels
and ‘Average of Return on Dollar Spent’ we see that each publisher has their own distinct
advantages. For example, customers looking to find western European destination results,
“flights to Florence” had a 1572% return when search via Google, with no such result on
Yahoo. Conversely, any dollar spent to link ‘European airlines’ with sales on Google was
a dollar-for-dollar waste, whereas Yahoo had a ROA of 1424%

2. How can campaigns be improved to increase overall value gained from investment
with a search engine publisher? Should keywords be added or dropped from the
campaign? Should campaign tactics or copy be adjusted to improve campaign
performance?
 To optimize a sponsored campaign, the campaign had to improve on one or more of the
following: cost-per-click reduction, increase in bookings, net revenue, and revenue per
transaction, return per transaction, and overall performance by engine, or other
performance metrics that could improve the net revenue of a particular campaign. We need
to understand how to maximize the net revenue and the ROA of its Internet marketing
campaigns by evaluating alternative strategies. We also constantly need to look to improve
our online performance through innovative thinking and testing new approaches. The level
of analysis and reporting we receive from them is crucial to the continuing success of our
business. It is essential, especially in an increasingly saturated airline industry and the
emergence of travel aggregators, that Air France target and pay for the keywords that are
most effective on each site. We can decipher from the data that geographically based
campaigns are sensitive to certain cities, thus destinations in France. This data not only
gives us insight into which regions frequent which French metropolises, but such data
should be further used to implement dynamic pricing. Pricing based on the city from which
the search took place, should correspond quite efficiently with the airport from which
someone departs on AirFrance. This should certainly be implemented one step deeper, on
a publisher-basis. Seeing as how different sites have their inherent advantages, based on
the order of the search result, or on the bid strategy, keywords (including city names and
words like “cheap” or “international”) should be exercised on a case-by-case
basis. Obviously this is an idealistic approach. However, to truly isolate those marketing
efforts that contribute to Air France’s web traffic and the facilitation of ticket sales, the
extra effort would certainly be worth the initial cost. By the estimates, only 322 out of
4,510 keyword searches, across the various platforms had positive return on the
investment. With overall returns being so high, across all websites, we can see that this
cost-model for advertising will sustainably remain effective. However, when bidding on
useless terms can be eliminated, we can further increase margins by not only saving on our
expenses when bidding for terms, but by increasing the costs for our competitors to
advertise.

3. What are the most important KPIs, and what impact will campaign changes have on
these KPIs?

The most important KPIs are the ‘average return on ad dollars spent’, the ‘total cost’ and
the ‘probability of booking’. Between these three key variables we can execute a balancing
act between margin benefits and overall expenses. We would want to remain flexible, given
the evolving technology platform, to the overall expenses accrued and the percentage return
we like to receive. The reason we are not deciding to prioritize average cost, is because
large expenses can be incurred if the high percentage rate accompanies such an
investment. Likewise, if something has slim margins but is relatively cheap, we see no
reason why not to maintain our presence on web, as people search, purchase, and later
recommend to others. ‘Probability of Booking’ is a reinforcement variable. We can do a
sensitivity analysis between overall investment in advertising and the rate of return on that
investment, but to ensure that little inefficiency are explored, we need to make sure that
 the investment has at least a minimal benefit. It can be debated whether exposing Air
France to the internet user, whether their cognizant or subconscious noticing Air France as
an option is worth the overall investment of advertising via a specific channel and whether
it has the potential to place Air France into minds of travelers as a prominent airline
provider.

4. How should future SEM campaigns be structured? In the past, Media Contacts had
concentrated on Google, Microsoft, and Yahoo; was there now an opportunity to
optimize search advertising with meta-search companies such as Kayak?

The future of Search Engine Marketing (SEM) campaigns should be structured with more
emphasis based on the Pay-Per-Click or Sponsored Search. This is because under Pay-Per-
Click, relevancy is determined by the match of keywords bid on by the advertiser, the
amount bid for those keywords in comparison with competitor bids, and the number of
times an advertisement had been clicked on for a keyword. This can increase the accuracy
of the results for its users based on what they are searching for. Furthermore, an advertiser
is only paid for an advertisement when it is clicked which can decrease the unnecessary
expenses for advertisers. Pay-Per-Click could also reduce with volume as the search engine
provider determine that the advertisement was more relevant to those keywords based on
the search results and the click of its users.