Allison WightmanAllison Wightman is head of e-business for Virgin Atlantic, with responsibility
for the global flagship, about a dozen associated sub-sites,
Commercial Insight and Database Marketing.

We talked about how the airline achieved a 40-to-1 ROI using attribution modelling to manage the online marketing campaigns and attribute
sales to the correct channel. The topic is one that will be addressed at the Econsultancy Peer Summit on June 2 in New York City.

What made you start?

We hadn’t done any detailed analysis around our online channel until about two
years ago. When we put some focus on it, we saw we had inaccurate reports and
digital advertising sales results that were adding up to greater than the web
channel delivered in sales.

We realized we needed something that would deliver one version of the truth and
we went with Tagman. Some of the other tagging solutions that had been around
for some time all fell short. We wanted to look at our results for natural
search as well, and Tagman was one of the only tools that did that.

How did you begin?

Initially we implemented the “last click wins” model, which was a change. Since
we’ve done that we’ve really been able to see clearly the path that customers
are taking towards conversion. In the future, we intend to extend this model to
what we are calling a “best click attribution” model where we reward
different forms of the marketing mix in different ways.

We want to reward based on the contribution the different parts of the media
mix are making to the overall goal. You find customers do different things
depending on whether they are in the consideration or the purchase phases.

Content partners are good at influencing, say, choice of destination. A meta
search engine may be about closing down a sale at the other end of the purchase
path, to do a comparison for best price. You don’t go to a search engine to
decide where you want to go on holiday.

The “best click attribution” model is something we plan to review
regularly at least twice a year, if not quarterly. We are currently in progress
of the analysis to come up with the initial model.

Can you articulate the factors that
go into the model?

In terms of what is going into it, it includes all of our online marketing
efforts — whether it’s display, affiliate, or whatever – they are all being tagged. Because all
this activity is in Tagman, we’re able to see what each contributed.

Its the level of detail we are now looking at which has moved our thinking on.
For example, if a customer’s journey was from email to display to paid search.
And then they booked. We’d want to know what media played a part?

Most people might say, “Oh yeah, we know what our CPA is for each part of the
media mix.” What we’re doing is going to a whole new level of detail. It was an acquisition
email, then a branded search and finally a retargeted banner.

Did they come
early in the path or later in the path and what weight do we apply to that?
We’ll be working with a partner to analyze that. No
tool can effectively analyze all the data for you as well.

Are there changes you’ve made already
based on what you’ve discovered?

Yes, part of our reason for implementing this was around duplicate commissions.

For the investment we made in Tagman, in a very short time it paid itself back
by eliminating double commissions. We’ve calculated a 40-1 ROI.

Before, for example, we would pay a commission to a direct partner search
engine, and we might also pay an affiliate. As a result, that’s why you end up
with fewer sales than the money paid out in commission indicates.

“Best click attribution” level takes our understanding to a whole new
level. Basically, in time we’ll be able to influence our budget allocation much
more effectively. Partners will be rewarded for the role they play in
conversion even if they’re not the last click.

Content partners, for example,
are likely to benefit as they are rarely the last click yet the effort is
greater to produce high quality pages that drive traffic.

Is this a global effort?

Yes. Changes within our marketing department happened at the same time where we
centralized most of the marketing planning activity to the UK and that’s made
it much easier. Marketing was centralized partly to reduce costs but also to
enable us to manage the brand more effectively.

These days we create assets
suitable for localization from the outset maximizing best practice on
successful campaigns globally.

Were there any big surprises that arose
out of your initial work?

Not that I’ve got an example of for you yet. We’re quite excited by the
potential though to use it as a lever to direct marketing dollars.

What role does online play in your
overall marketing?

It’s a key part of everything the airline is doing. We’re not focused on
driving people down one channel or another, we believe the customer should have
a choice of all channels
at any particular time to meet their needs.

With all the emerging platforms — mobile, web TV, tablets — we’re confident
that with the universal tag we’ll be able to manage all activity effectively as they
are quite progressive in the way they are developing the product and we’re
hoping to work closely with them to innovate in this area. 

How does universal tagging affect
download times?

We have over 100 campaigns running globally at any one time. If you’ve got lots
of tags from different vendors all sitting on the purchase page, it starts to
become a very slow page. Tagman can handle all that. It shows one piece of code, which
makes for a much quicker download and improved conversion.