April 21, 2003
How To

Cross-Media Metrics: How Radio Ads Affect Site & Search Marketing Results - Data from esurance

SUMMARY: How does offline marketing affect online sales? Can online and
offline campaigns each increase the other's impact?

Everyone knows that offline and online have some relation to each
other. The problem has been measuring it. Hear how the
marketers at esurance tracked cross-media metrics, and what their
results were. (Yes, regional radio ads can work.)
Two years ago, Tolithia Kornweibel, Marketing Manager for
esurance, began a cross-media regional radio campaign, spending
the bulk of her marketing budget in online advertising while
testing geo-targeted offline campaigns in a way that allowed her
to track all incoming traffic.

That tracking ability enables her to change her campaigns on a
monthly basis in order to target the most effective regions.

Results: New customer acquisition numbers have grown by 30%,
with acquisition costs decreasing by about 10% each quarter
since the cross-media campaign began.

The quality of customers has improved. “It’s not just that
more people respond,” Kornweibel says, “it’s that more motivated
people respond. People who come to us online are more likely to
be from smaller insurance companies, looking for a cheaper

We talked with Kornweibel about how she and her team applied
their online tracking methodology to offline advertising.

-> In the beginning: Tracking online marketing

Originally, the majority of esurance’s marketing budget was spent
online, on paid placement campaigns, paid inclusion campaigns,
and category marketing such as ads on the Insurance Center

“At that point, our ROI was imminently trackable, through the
data we collect on our site, and to a lesser degree through ad-
serving tools like Atlas,” says Kornweibel.

The site, designed like a “data collection funnel,” allows them
to track conversion metrics and to segment customers from the
different online sources. By tracking URLs (for online
advertising) and referring URL (for search engine optimization),
she could see where customers were coming from and could
determine effectiveness of her online campaigns.

While they still spend approximately 70% of their marketing
budget on online campaigns, “We wanted to be able to apply the
same methodology to offline advertising,” she says.

-> Tracking local radio ad results online:

Kornweibel’s team began testing geo-targeted radio campaigns,
with markets grouped into eight different geographic categories.
Then she followed the untracked traffic (people who came directly
to the site by entering the Web address rather than finding it
via a search engine) to judge the success of the campaign.

The first thing potential customers must do on the site is to
enter their zip codes. The ad-serving tool Atlas shows the time
that the consumer visited. These two pieces of information (the
where and the when) allow Kornweibel to see if traffic is
increasing at times when radio ads are running in various
geographical locations.

“So with every person who comes to our site and achieves an
important milestone like getting a quote or buying a policy, we
look at the media source, the time, and the geography,” she says.

Looking at the success of each geographical category, they were
then able to come up with recipe for each market (i.e. this
market will do better with this number of rating points a week,
with talk news and adult contemporary as the best format.)

-> Cross-media analysis: How radio affects search marketing

“When running offline campaigns in radio, we see a lift in
referral traffic from search engine marketing,” says Kornweibel.
“We see anywhere from 20 to 40 percent in a particular market
when we’re doing an offline campaign in the same market.”

Knowing where each customer comes from allows Kornweibel to come
up with a cost to acquire that is more specific than a general
acquisition cost.

“Because we are a planned purchase and it takes a couple of weeks
to make a decision, we look at the media that brings people to
the site both for their first visit and for their subsequent
visits,” she says.

She looks at all of a customer’s milestones that lead to a
purchase. “When a conversion occurs, we attribute it back to the
first visit and whatever media source that first visit was
created by, but then we also look at the interim to see what
medium created those milestones.”

For example, she says, they may run a certain number of points on
the radio in a certain geographic area. If she sees a significant
lift in the number of people searching for them on search
engines, some of the spend for search marketing she will actually
account for as supportive for offline media.

“We try to do that with everything so that it fits into one
holistic picture,” Kornweibel says. This helps them figure out
which program has too much sunken cost.

“If I’m spending ten percent on a particular program and measure
only from cost per impression or cost per visit or cost per
purchase, I’m missing the fact that some of the budget for that
program is going to be supporting other programs,” she says.

Of course, that is where it gets a little complicated, there is no
actual formula they use to come up with what percentage of a
program goes to support another program.

“It’s very data driven, not algorithmic,” she says. “Ultimately,
no media program is able to be way off the ROI metric.” However,
she says, they recognize that, while some programs look more
expensive, they are adding additional value to other programs.

Currently, between 30% and 50% of her budget is spent offline.
“That’s based, at the end of the day, on getting to the right
overall acquisition cost for new customers: total media spent
divided by total number of policies.”

-> Tracking customer lifetime ROI metrics:

Once someone purchases, Kornweibel’s team keeps a running tally
of the premium they are paying and the profitability metrics
(retention, payment, claims) which they track back to the various
media, as well.

“That helps us plan our budget, becoming more efficient,” she
says. “With certain programs, the conversion metrics may look
low, but we’ve saved the program because we’re getting a better
quality of customer. Others we’ve saved because we may spend more
in search engine marketing but we’ve gotten back sunken costs in
other areas.”

Her team evaluates the CLTV (customer lifetime value) through the
data collected on the site: Demographic and credit information,
driving record and insurance record information.

Kornweibel says she learned these four lessons along the way:

Lesson #1. Start with infrastructure

Make the tracking infrastructure work the right way. Come up with
the theories behind how you’re going to evaluate things.

“We spent a lot of energy since we began developing our real-time
reporting tools,” she says. “Since we’re in insurance, we have a
lot of nerds. In-house analytics are incredibly important—you
need someone to make sense of the numbers.”

Lesson #2. Flexible media buying

Have flexible media plans that you can change them within a
matter of days or weeks, not months. Kornweibel’s team looks at
conversion metrics for every single media program every day.

Monthly, they look at milestone analyses and interim visits, and
segment their customers behaviorally, then re-do the budget all
over again.

“We’re small enough that when we do these budgets, we can all fit
in a room and haggle over it,” she says. “It’s a hassle, but
that’s been the hallmark of our success, especially in these

Lesson #3. Feedback loop to media buyers

Have a feedback loop from whoever is measuring the business
success goals of the advertising strategy back to the media
planners on at least a monthly basis.

“In my case, I need to be able to analyze the programs from the
quality of customer. I need to know how much premium I’m
getting and how much loss against the premium. So I need to be in
touch with whoever’s looking at the numbers.”

Lesson #4. Customer vs. transactional-centric databases

When it comes to your tracking infrastructure, have a consumer-
centric database rather than a transactionally-measured database.

It is all quantifiable, Kornweibel says. “We can say that search
engine marketing might cost one dollar, but of that dollar, 25
cents goes to recoup sunken costs of other media, like radio.
We’re running maybe 30% offline and 70% online, but they’re both
generating efficiencies.”

Where does she think she is getting the biggest bang for her
marketing buck? “We don’t look at the programs individually. As
long as we identify positive combinations and keep acquisitions
costs down, we can keep getting a bigger budget.”

Want to see Kornweibel in person? She will be at the next ADTECH
show in San Francisco, June 16th-18th. Info at

Improve Your Marketing

Join our thousands of weekly case study readers.

Enter your email below to receive MarketingSherpa news, updates, and promotions:

Note: Already a subscriber? Want to add a subscription?
Click Here to Manage Subscriptions