Display Campaign Optimizer: Some Recommendations from a Beta User
Google just announced that the Display Campaign Optimizer (DCO) is open for everyone. Before you start jus using it, I wanted to give some feedback on how its worked for a few accounts as I’ve been using this feature for more than a year.
Please note, these are large accounts that participated in the beta. Some of them have spent more than $40 million on DCO all time with monthly budgets of a few million just for DCO. Other campaigns have spent somewhere in the $5-$10 million range. I think the smallest campaign was spending around $10,000/month on DCO. So, there was a lot of data available.
First, you should know how successful display campaigns are run, and then we’ll look at the differences of how DCO works.
Successful Display Campaigns
Most successful display campaigns use a few steps in optimizing their overall performance:
- Choose a handful of keywords by ad group
- Let the ad groups collect data
- Examine the placement data
- If a site does not meet your goals – block it
- If a site meets you goals – make it a management placement (often in a new campaign)
As your goal is to find sites that meet your target CPAs, and you’re constantly playing with ads, bids, and exclusions and that can be quite time consuming.
Note: Another way to manage display is to just use managed placements, which is usually best for small budgets.
The main question is: can DCO take out some of this work? Let look at how DCO has worked.
How Display Campaign Optimizer Works
With DCO, you only write ads in the campaigns and set a CPA. You do not choose keywords, placements, topics, interests, or other targeting options.
Based upon your ads and landing pages – Google does all the work.
So, what you lose is control. What you gain is time.
The Big Change to DCO: Minimum Conversions
During the beta, you needed at least 150 conversions a month to be eligible for DCO. Google just opened it to everyone as they lowered this threshold to 15.
This is similar to CPA bidding. When it launched you needed 60 conversions/month. Then it was lowered to 30, and finally 15. While CPA bidding is wonderful if you have 60-100 or more conversions each month in your campaign, it does not work well if you have 15 conversions and 10,000 keywords. There just aren’t enough data points for Google to optimize.
Because you set a CPA, you must be using the Conversion Optimizer bid system with target CPAs enabled.
Should You use DCO?
If you have a large budget and are willing to try DCO out with a budget of at least $5,000/month AND you have found some success on the display network; then please give it a try – you’ll probably see some good results.
If you can spend at least $10,000-$100,000/month on a DCO campaign and have had some success on display – then definitely try it.
If you have a budget that is only going to receive 15-50 conversions a month on the display network, you can either:
- Proceed with a lot of caution as it really might not work
- Use DoubleClick adPlanner to find managed placements and maintain lots of control
You are giving up control. So if you have a well organized display campaign that is working well and you don’t have ‘new budget’ to add to DCO – meaning you’ll be moving existing budget – then proceed with caution as you’re pulling budget away from effective advertising.
I’ve found more success with DCO when you give it a new budget and then as it works, you can transition budgets from existing display campaigns that are working marginally to the DCO campaigns.
How to Enable Display Campaign Optimizer
Enabling DCO is easy. Navigate to the campaign settings and then at the bottom of the page, change the targeting options as seen in this screenshot.
The Confusing Part – You Can’t have Keywords, Placements, or other Targeting Options so How Do You Get to 15 Conversions?
In the beta test you needed a rep to enable this for you. So, you or your rep made the campaigns, you added your CPAs, budgets, and ads and then the system went to work.
You could not have keywords, placements, topics, interests, or other targeting options enabled.
Now, in the new system you must have at least 15 conversions – which means you must start with placements, keywords, topics or some targeting option that will get you enough conversions so you can enable CPA bidding and DCO.
So, it would seem that once you meet the threshold you have to delete all your targeting options for it to work or pause your old ad groups with targeting options, make new ones, and then let it go to work.
This step seems very odd at the moment. I reached out to Google to see how this is going to work. If anyone knows, please leave details in the comments.
In the beta, you really needed to give the tool a couple weeks to collect data and really get going. Sometimes it worked immediately, other times it took a while. Most of those reasons were based upon Google collecting enough data to show your ads. As the new system requires a campaign that already has conversion data, the wait period for success or failure determination could be much lower.
So far, I have put several accounts into the beta. In almost all cases (there was one exception) DCO did a great job of finding new conversions and acceptable CPAs.
It did fail for one account; but I have a feeling that part of that issue was it was a product with a landing page that didn’t have lots of text data on it (it was images and video) and DCO mostly had the ad copy to work from for targeting purposes.
As DCO uses only ad text and landing page data to serve your ads, landing pages with at least a few paragraphs of descriptive text seem to do a little bit better overall – at least to start. Once the system gets going, it has enough historical ad serving data that the landing page info matters a bit less.
The landing page importance of DCO for small to mid sized budgets that won’t have lots of conversion data might (again, I don’t know as there haven’t been small accounts in DCO yet) really come into play for proper ad serving. In other words, if DCO isn’t working and your landing page doesn’t have much text, changing the text of your landing page might help DCO serve your ads better.
If you have additional budget that you aren’t spending and have found success on display – definitely try it out.
If you have had marginal success on display and a healthy budget, then put some budget into DCO to see if it outperforms what you have been doing.
If you have a small to mid-sized account and are an experimenter – try it out – it could be really good (or not).
If you have a small to mid-sized account and prefer to let others fail or succeed first so you can make a highly informed decision about trying it out – then wait until some more results from these types of accounts are published.
Overall, I’ve really enjoyed it – but I have been working with large accounts with lots of data. If DCO works for small to mid-sized accounts, then the display network just got a whole lot more attractive.
- DCO announcement from Google
- DCO help page (note, its not updated to reflect the fact its open to all)
- Google Promo video for DCO with Dale Carnegie
Google video on Display Campaign Optimizer: