With the above said, we were set the following targets:

• Increase conversion volume
• Decrease cost per acquisition (CPA) by focusing on creating the image of Ehrlich as a “local” pest control brand
• Reduce non-brand cost per acquisition (CPA) by 68%

To achieve the objectives set by Rentokil, we built out a campaign structure that focused on showing users the most relevant possible location information and phone numbers in ads across desktop and mobile devices.

Specific campaigns were created to individually target as many of Rentokil’s key cities as possible – serving users searching in these cities with a local phone number to call, as well as stressing Ehrlich’s expertise in that city.

Due to the size of Ehrlich’s business, the implementation of this strategy was the complicated bit.

We worked out that to target each city with relevant ad copy we would have to generate 25,000 different ads to cater to 3,000 different locations. No small feat – and a job which would take weeks’ worth of man hours.

Step 1: Bespoke tool

To deal with this problem we created a bespoke tool which measured the number of characters in each combination of pest and location. Based on this number the tool would insert that information into one of a number of ad copy variables (which changed based on how much of the ad space was taken up by the pest and location).

The tool then looked at the location these ads were targeting and pulled in the phone number from the closest local Ehrlich office as a call extension.

This system allowed us to create highly specific ads, combined with highly specific phone numbers at an incredibly large scale in a matter hours!

This ad structure was especially successful for mobile users – where the increased likelihood of a call directly from the ad means appearing relevant and local is all the more important.

Based on our initial success with this location strategy– we developed the second arm to our plan, which aimed to group lower volume areas together in a way that still gave users a highly localised experience.

Step 2: Location-based targeting

This second part of the strategy drew a 50 mile radius around each of Ehrlich’s 60 key office locations.

We then stripped the name of each town that fell within that radius, and generated a master list of pest control related keywords containing those town names.

We linked these to ads whose main text would promote the local office and a link to that office number – with a headline that was dynamically generated based on which town the user was searching from.

We then made sure that the local experience continued on site, but rather than getting Rentokil to develop hundreds of location specific landing pages – we worked with them to develop a dynamic system. This identified where a user was searching from, pulling in their location dynamically to the header of the landing page.

This holistic local response, from ad copy and phone number to landing page proved very effective over the course of the campaign.

Results

The campaign has been a resounding success – so much so that Google is using it as an example of the value of highly localised mobile ready ads in upcoming internal and external presentations.

• 203% increase in conversions
• 238% increase in direct calls from PPC ads
• 56% increase in non-brand (generic) conversion rate
• 74% decrease in non-brand (generic) CPA
• 47% increase in mobile conversion rate

“In the second half of 2014, after the high season ended, JCEhrlich had a strong focus on developing high volumes of pest control leads at an efficient level. We worked closely with our agency Merkle, and took an aggressive approach to the optimization of our campaigns, both at a geo-targeted and pest species level across all our serviceable markets. We met and exceeded our expectations and KPI's on delivering a substantial increase in year on year leads to the sales organisation. Aligning our Paid Search Campaigns for maximum conversion this way helped beat our efficiency goal while driving incremental revenue.”

Randolph Carter, VP Marketing, North America