AdWords
4.3K members online now
4.3K members online now
Improve your AdWords performance and boost your ROI, CTR, and Quality Score
Guide Me
star_border
Reply

Campaign Draft Experiments Good for Testing Many Major Changes at Once?

[ Edited ]
Visitor ✭ ✭ ✭
# 1
Visitor ✭ ✭ ✭

Just starting to help a client with an 14 year old account. Many campaigns, but one in particular is extremely important to the success of the business.
This campaign was quite profitable many years ago but has been through several managers and performance has gradually dropped through the years. It is now just barely profitable.

Looking around in the campaign I can see that it's just a flat out mess. Amazing that it's even doing as well as it is. It's a very large campaign with many ad groups and ads.
All the while looking at this mess I'm thinking it would be far easier to just build out a brand new campaign then try to improve this one.
However, I cannot do anything that would risk this crucial revenue stream for the client. Even though it's barely profitable, I cannot risk moving revenue backward before moving forward.

So... I see this new Campaign Draft Experiment feature. And I wonder if it's really meant for testing small incremental changes? Or would it also be wise to use the draft experiment to test a completely rebuilt draft campaign against the old barely profitable campaign?

Thanks much for any advice...

Campaign Draft Experiments Good for Testing Many Major Changes at Once?

[ Edited ]
Participant ✭ ✭ ✭
# 2
Participant ✭ ✭ ✭

Assuming the campaign has conversion tracking in place, and has had it in place for some time, the historical data is the key. What has worked for me in that situation is to begin by creating a performance baseline for the campaign, summarizing the key performance indicators (usually ctr, avg cpc, impr/click volume, conversion rates, cpa's, etc) over a long time frame, and then also a shorter term time frame (e.g. year to date, or 90 days) in a spreadsheet. 

 

At the start of the baseline process I also make a careful study of every setting and configuration option, from ad delivery to geo targeting, ad extensions and so forth.  Often I will find configuration errors or oversights which can then be changed and tested in isolation. 


Quality score is not a historical parameter however when starting work on a campaign, it is also good to sort the keyword base by quality score, then create a "quality score distribution" chart on a scale of 1 to 10.  Poor quality score keywords are often useful areas to focus on. 

 

Then, try and identify where ad spend is being wasted and start with (reversible) changes which begin to take wasted ad spend out of the campaign.  Look at performance by ad group and keyword, gradually shutting down things which are consuming ad spend but are not delivering conversions.  As you go through that process and the campaign starts to become leaner, the workhorse ad groups and keywords which are delivering results will start to become clear.  

 

Once the campaign is "leaned down" and the productive parts of it start to become clear, then opportunities for enhancement will begin to become obvious. 

 

The adage, "first, do no harm" can be your best friend.  And always change as little as possible at one time, so you can observe the effect of changes in isolation, to the extent that is possible. 

 

Sometimes working on successful campaigns like that, even though they are marginally successful, can feel a bit like overhauling an airplane while it is in flight.  So best of luck. Obviously making sure the client is "bought in" for possible short-term setbacks is important. 

 

Periodic backups of the campaign using Google AdWords Editor can also be helpful, in case you need to restore the campaign to it's state from a day ago, week ago, etc. 

 

hth; not the type of project for the faint of heartSmiley Happy

 

Re: Campaign Draft Experiments Good for Testing Many Major Changes at Once?

[ Edited ]
Badged Google Partner
# 3
Badged Google Partner

Hello @Tj F

Drafts and experiments let you propose and test changes to your Search and Display Network campaigns. You can use drafts to prepare multiple changes to a campaign. From there, you can either apply your draft’s changes back to the original campaign or use your draft to create an experiment. Experiments help you measure your results to understand the impact of your changes before you apply them to a campaign.

This post will explains how campaign drafts and experiments work so you can decide whether they fit into your overall advertising strategy.


Example

Anthony helps run the advertising efforts for a medium-sized clothing company. His boss suggests that he try a new strategy for their search ads, and he'd like Anthony to show him the proposed changes before they are put into effect. Anthony creates a draft of an existing campaign in his company’s AdWords account. The draft lets him make changes at the campaign level before applying them to a campaign. After Anthony finishes making his changes in the draft, he shows them to his boss, who approves them. From there, he applies the draft back to the original campaign.

Next quarter, Anthony and his boss decide they want to change the bids for their campaign but they want to be confident these changes will improve performance. Anthony creates another draft with the bid changes and runs a month long experiment to test these changes against the original campaign. After measuring the experiment’s results at the end of the month, Anthony finds that his bid changes performed positively, so he applies his experiment to the original campaign.

 

How campaign drafts work

Drafts let you prepare multiple changes to a campaign without impacting its performance. When you create a draft, you’re mirroring your campaign’s setup. From there, you can make updates to your draft just as you would in a normal campaign. At any point, you can leave and return to your draft to make additional changes to it, or discard the draft altogether.

After you’ve finished drafting your changes, you can apply your draft to the original campaign or create an experiment to test how your changes perform against your original campaign.

Some features and reports aren't available for drafts. These include the following:

  • Dimensions tab
  • Ad schedule report
  • Category & Search terms
  • Auction Insights
  • Display Placements report
  • Scheduled email reports
  • Bid landscapes
  • Ad customizers that use "Target campaign" or "Target ad group"
  • AdWords Campaign Experiments
  • Keyword diagnosis
  • Some automated bid strategies:
    • Target search page location
    • Target outranking share
    • Target return on ad spend (ROAS)

 

How campaign experiments work

After you’ve finished a draft, instead of applying your changes to your original campaign, you can convert your draft to an experiment. As you set up your experiment, you can specify how long you’d like it to run and how much of your original campaign’s traffic (and budget) you’d like it to use.

When a potential customer performs a search on Google or a search partner website, or loads a webpage on the Display Network we'll randomly make either your original campaign or your experiment active for the auction, depending on how you’ve split the traffic share between your campaign and your experiment.

As your experiment runs, you can monitor and compare its performance against your original campaign or change the dates of your experiment to end it early. If your experiment performs better than your original campaign, you may consider applying your experiment to the original campaign. You also have the option of converting your experiment into a new campaign with the same dates and budget as your original campaign and pausing your original campaign. Note that while you can have multiple drafts for a given campaign, only one of those drafts can run as an experiment at a time.

 

Features that aren’t supported by experiments

Experiments generally support the same features as campaigns, with a few exceptions:

  • AdWords Campaign Experiments
  • Ad customizers that use "Target campaign" or "Target ad group"
  • Some automated strategies:
    • Target search page location
    • Target outranking share
    • Target return on ad spend (ROAS)

Campaign Draft Experiments Good for Testing Many Major Changes at Once?

Visitor ✭ ✭ ✭
# 4
Visitor ✭ ✭ ✭

Thank you kindly for the thoughtful and detailed response Steve... 
I will proceed accordingly.
At the end of of this overhaul, I suspect the campaign will probably look about the same as if I had built it up from scratch using all current best practices. 
So I was hoping to circumvent the lengthy overhaul process, and just create a whole new campaign using the draft feature. 
I would still have to test of course, but I have to think that a brand new well built campaign would quickly begin to outperform the old sickly beast I am looking at. Then I could safely just apply all the changes at once to the original campaign. 
But I guess this may be much more than campaign draft experiments are meant to do. 

 

Campaign Draft Experiments Good for Testing Many Major Changes at Once?

Participant ✭ ✭ ✭
# 5
Participant ✭ ✭ ✭

That's another way to approach it... just build a new campaign using all best practices.  Then shift say, 20-25% of the ad spend to the new campaign and as you dial it in, it starts to outperform the old campaign, shift more and more of the ad spend to it. 

 

Might be more time efficient.  However I find there is great benefit in studying the campaigns with some mileage on them as they usually wield some less-than-obvious secrets, both good and badSmiley Happy Which can be either further leveraged, removed, or applied in a new campaign.