Lower Ad Spend, Steady Conversions

  • October 23, 2018

Milliyet Emlak Case Study – Overview

About Milliyet Emlak: Established in 2005, Milliyet Emlak was the first real estate listing website in Turkey. It features various house listings, construction updates, upcoming housing projects, and so much more – covering pretty much everything in the real estate sector on its website. After well over a decade in this business, the company has grown and innovated to adjust to the ever-evolving digital world to enhance customer satisfaction.

Even today, it is one of the most customer-centric companies in Turkey; it acts as a matchmaker between people and their dream homes.

Objectives: Even though Milliyet Emlak already had a well-implemented ad campaign, we could envision a plan that improved on something good – to make it even better. Our primary aim for this campaign involved reducing the company’s ad spend without affecting conversion. In this case, the ad spend reduction relied almost wholly upon reducing CPA (Cost Per Acquisition) and CPC (Cost Per Click) metrics.

Final Outcome: Our optimization efforts brought about a whopping 26.8% reduction in CPC and an admirable 17.6% reduction in CPA in the span of a mere 3 months. That’s not all. These metrics are on a steady downward trend even today – continuously reducing the company’s ad spend by a significant margin without affecting conversions.

Challenges

Ad spend is more or less linked to conversion rates directly. Increasing ad spend increases impression rates, which invariably boosts the conversion metric. Conversely, reducing the ad spend usually causes a dip in conversion as well.

Our biggest challenge was reducing CPC (Cost Per Click) and CPA (Cost Per Acquisition) without affecting the current conversion in any way.

Our Hypothesis

After a careful study of the marketplace and end-user behavior, we hypothesized the following –

CPA can be reduced without reducing conversion by tweaking the CPC according to User Behavior.

For example, users who visit Milliyet Emlak from a mobile device during work hours usually look to research before they are ready to take action. On the other hand, a user who has visited the website, say, four times in the past week is far more likely to take action.

Our Approach to the Problem

Since digital footprints can be used to track user behavior reliably, the team diligently analyzed and categorized the behavior of Milliyet Emlak users. We, then, proceeded to optimize the campaign by splitting the ongoing ad campaign into multiple, optimized ad campaigns. We used the following factors to organize, target, and split them –

  • Device (Mobile – Desktop and Tablet): Since we found a significant disparity between the conversion rates on different platforms, the Boost team started by bifurcating mobile and desktop into discrete ad campaigns. The analysis that showed significant behavioral differences between these two sets of users made this decision fairly straightforward.
  • User Location: The objectives, intentions, and behavior of users varies by location. In Milliyet Emlak’s case, this is partly due to non-uniform brand recognition and partly due to location-specific services/listings on offer. Keeping this in mind, the Boost team customized and split ad campaigns by city.
  • User Search Query: All search terms cannot be treated equally and our analysis showed vast differences in conversion, depending on the search terms used by visitors. Since generic search queries had much lower conversion rates compared to more specific ones, we split campaigns by search query as well.

Once we were done re-organizing the Milliyet Emlak ad campaign, we ended up with over 404 campaigns and 10603 discrete ad groups – allowing us much more scope for reducing costs.  Further analysis led us to 7 remarketing lists, which featured unique user behavior. The users in these lists behaved differently depending on the time of day they visited the website.

With this information in hand, all we needed was a way to modify CPC on the fly – changing on an hourly basis. We used the Optmzr tool to run a custom Google Adwords script that allowed us to do just that.

Technical Challenges

Even though we knew exactly what to do, how to go about accomplishing it was a challenge in itself. We needed a convenient way of making frequent changes to the CPC.

To tackle this problem, the Boost Team created a self-updating dashboard to see real-time performance metrics on Google Spreadsheets using the Google Analytics API. Not only could it be used to see the performance of the Milliyet Emlak ad campaign, but the dashboard could also push changes to the Adwords editor by leveraging the Adwords API – giving easy, yet comprehensive control of the ad campaign to the company.

This dashboard automates many manual processes and reduces the time spent on these tasks – giving the company a boost in efficiency in addition to boosting its return on ad spend.

Results

In a span of three short months, the Boost team delivered what it had set out to do – reduce the CPC and CPA without affecting conversions – effectively reducing ad spend without compromising on sales and conversions.

Cost Per Click (CPC) Reduction

The Cost Per Click, or CPC, is the rate charged by advertisers every time someone clicks an ad. Hence, lower the CPC, lower the cost to your company.

We saw a 26.8% reduction in CPC over a period of 3 months, with a steady downward trend.

Cost Per Acquisition/Action (CPA) Reduction

Cost Per Acquisition/Action, or CPA, is the rate charged for a click AND a specific action (a purchase, sign up, and so on). Similar to CPC, the lower the CPA, the lower the ad cost to the company.

We noticed a marked reduction in CPA over the 3-month period. The metric reduced by 17.6% and is still trending downward.

Leave a Reply

Your email address will not be published. Required fields are marked *