Increasing ROI by Applying Modern Portfolio Theory to Pay Per Click Optimization
This “Budget Awesomization” strategy is based on modern portfolio theory. And that makes sense because managing pay per click spend is not unlike investment management – http://en.wikipedia.org/wiki/Modern_portfolio_theory
The lesson? Invest more ad spend in PPC campaigns that produce greater ROI. Here’s how:
- As usual, pause campaigns that don’t meet performance targets.
- Invest in campaigns that meet minimum performance threshold, but only invest enough to ensure sufficient results for performance-increasing tests.
- Invest most of your budget in the highest ROI campaigns.
How to Prepare Your Account Structure for Budget Optimization
To maximize returns, we must put AdGroups into campaigns according to their ROI. Since we can only control budget at the campaign level, if we want to vary how much we spend based on the AdGroups’ ROI, we have to put the highest ROI AdGroups together, the lowest ROI AdGroups together, and so on.
[A note about ROI and ROAS: I use ROI generally to mean return on investment. ROAS is a specific formula also called value/cost in AdWords reports. ROAS is a percentage that equals revenue divided by ad spend. A 500% ROAS is the same as spending 20% of your overall revenue.]
An Awesome Example of Pay Per Click Budget Awesomization
At Fuel Interactive, we were running a PPC account for a group of 12 hotels. We were competing against another optimization company, and had a few limitations placed on us: we were only allowed to send traffic to a lower converting multi-hotel website and not to the individual hotel property sites. In many cases, the hotel property sites were converting at a higher rate than the multi-hotel site.
After a few months of initial optimizations, we grouped these hotel AdGroups into low, high, and very high ROAS campaigns. The client wanted maximum overall ROAS and didn’t require hotels to receive the same advertising spend, so we could optimize the account by spending more on the hotels that produced greater ROI.
Here are the groups we created, their definition according to ROAS level, and the percentage of the budget we allocated to each.
|Group||< Low ROAS||High ROAS||Very High ROAS|
|Definition||<500%||500% – 1000%||>1000%|
The proportions you define may have to start with what the minimum spend you need is for the lowest ROAS campaigns. For example, if you have six AdGroups in the lowest ROAS group, with a total of 12 ads, you want to get 100 clicks per ad per month to better evaluate your ad tests, and your CPC is $2, then you need to budget $2400 to that campaign. Adjust the other campaigns accordingly, or reduce the number of tests you’re running in that lower ROAS group.
The Problem of Underperforming Properties
Sometimes campaigns with lower ROI suffer from issues a PPC specialist can’t fix. For example:
- Product conversion rate is low
- Product has less compelling offer
- Product has reputation problems
Anything that inhibits conversion rate and clickthrough rate decreases PPC ROI.
If a client insists all offerings receive equal spend, poorly performing offerings drag down overall PPC performance. But clients who can be flexible on the allocation of ad spend allow for PPC ROI optimization while the website, promos, or reputation of the offering is improved.
The Awesome Results of PPC Budget Awesomization
In our example, we were able to achieve 530% ROAS even in the off season and during a recession. This success led the client to decide to move all their PPC spend to us, as well as some of their marketing budget from other channels that either can’t be tracked as clearly, or did not perform as well. We quadrupled the client’s PPC spend by demonstrating this level of ROI.
We now have the whole budget. There will be pressure to ensure that the higher spend maintains good ROAS levels, but the good news is we’ll be able to send traffic to whichever site converts the best for each hotel, and that will help with ROAS. We’re also going to test an interesting strategy I’ll write about more later: different ROAS targets based on customer lifetime values that vary by the prospect’s geographic origin.
Possible Complications During PPC Budget Awesomization
If you change nothing else, the overall account ROAS should improve. However, other factors may come into play: other ad tests you’re running, seasonality, the health of the economy, etc.
A funny thing happened on the way to the budget optimization: some of our ad tests in the low ROAS campaign were more successful than those in the High ROAS campaign, so the low ROAS campaign began to outperform the High ROAS campaign. This simply required shifting the better AdGroups to the High ROAS campaign and vice versa, but it was a fun topic to explain in the client meeting why the Low ROAS campaign had a higher ROAS than the High ROAS one ;-)
Brian Carter is the Director of Search Engine Marketing for Fuel Interactive, an interactive marketing agency in Myrtle Beach, South Carolina. He is responsible for the SEO, PPC, SMM, and ORM programs at Fuel and its partner traditional agency Brandon Advertising & PR.