Small Business Series #9 - Evaluating and Optimizing Campaign Performance
OK, so you’ve built creative, deployed campaigns and tested your messaging, and maybe even made your first sale. Congratulations! But just like catching your first wave, you’re back on the beach now and need to paddle back out to the lineup to catch another wave. How do you evaluate performance? How do you capitalize on your wins and mitigate your losses?
Well, there’s this slick thing that digital ads have called “Data” that lets you optimize your campaigns. Let’s look at a few ways we can review performance and make sure we continue to get better results from our campaigns. Effectively, these are the specific tactics of our “Battle Ship Method” from one of our original blog posts.
Here’s the TL;DR and high-level concept: Spend more money where you’re getting the results you want, spend less where you are getting less results.
Of course, it’s a little more complex than that so we’ll look at the specifics here more carefully in 2 primary channels, Search (Google & Microsoft Ads) and Social.
Reviewing Performance on Search Campaigns
OK, so you’ve read our SBS#3 (How to Choose a Paid Ads Channel) and have had success in a few of the campaign types (Brand, Nonbrand, Performance Max, Display?) and have started to see some results.
So keeping in mind that we want to reduce spending ad dollars where we aren’t getting results and redirect those funds to areas where we are, let’s look at the potential dimensions that we can segment
Reallocate on Account Structure:
So ideally your account structure is built in such a way that it groups ad groups and keywords in similar themes. This means that you can look at performance at the campaign, ad group and keyword level to allocate ad dollars appropriately. Since you need more data (generally clicks) to make a decision, you can make more frequent decisions at the Campaign level than you can at the ad group or keyword level.
For example, you can review the ROI of a given campaign and adjust the budget to that campaign fairly regularly, but you probably want to let a good amount of data roll through that campaign so that a single keyword within it has enough data in it to decide if you should keep paying for it or not.
If you’re using a lot of phrase and broad match keywords, make sure to take a peek in the Search Query report and set negative keywords for queries that are irrelevant to your business, or simply perform poorly.
Reallocate Based on Geo, Device, Placement:
Google can give you some interesting results in the many reports that are available. Many of these segments are only available at the campaign level (Geo, for example) so make sure that your campaign structuring is built in a relevant way.
If you notice that you’re getting better performance from a particular region, you can set bid multipliers to make sure you’re spending additional ad dollars in those regions. Conversely, if a region performs poorly for whatever reason, you can adjust a bid down so that you don’t spend as much. If things get really hairy, you can exclude regions altogether.
Placements are generally more for Display Campaigns, so you can audit where the ads were served. These reports can get pretty hairy with thousands of placements. Sort the report by impressions descending and make a pass a the highest volume placements to maximize the effects of your time managing the campaign.
Reviewing Performance on Paid Social Campaigns
Paid social is always conceptually simpler than Search, so this is fairly straightforward. If your account structure is built out in a way that makes it easy to see the performance of different audiences, you should be able to see fairly clearly which are getting you results and which are not. Typically, we like to break up performance at the placement level for ad groups (Instagram, Facebook, Messenger, etc…). Review the performance of each audience and channel and reallocate your budget accordingly.
What’s a bit more tricky is evaluating ad performance. Since there’s not a one-to-many relationship between ads and audiences, we typically deploy duplicate ads to all appropriate campaigns.
Flip over to the “Ads” Tab of Facebook Ads Manager and filter down to the name of the ad in question and evaluate the performance of that ad in aggregate against all campaigns and ad groups it’s deployed into. You may see that certain ads don’t perform as well with certain specific audiences so it may make sense to only shut down those ads within those audiences, but you may also see that a particular ad simply performs poorly relative to your other ad creatives so you can just shut them down en masse.
Am I Reaching Market Saturation?
One last consideration is that it’s quite possible to have reached all of a particular customer that you're trying to message in a campaign. For example, if you’re running a nonbrand search campaign, there are only so many people in a given month that are searching for “paisley printed silk pocket squares”. For this reason, it’s a good idea to keep an eye on a key search metric called “Impression Share” which is basically: of the number of people searching for a particular query; what proportion of those users have seen an ad from me?
If you’re reaching 80-90% of users for a given query and aren’t using Broad Match, you’re probably good to go and you’re not going to reach too many more users, even with substantial increases to your budget and/or bids.
In Social, keep an eye out for the “Frequency” metric. This is the number of impressions served, divided by the number of unique user accounts reached. Keeping this number below a 2.0 on a 7-day lookback is where we see the best performance for a prospecting audience and we see dropoffs in performance at a 3.0 for retargeting.
These benchmarks are from our experience with eCommerce clients selling pretty typical consumer goods, so if your product has a much higher consideration phase ( Higher AOV, bigger commitment of time like a vacation, etc…. ) then consider going a little higher. Also, consider that the number is an average so if you have 2.0 frequency, there will be some users who have seen the ad only once and some that will have seen it 3 or 4 times.
Conclusion
Reviewing ad performance is something that should be done at a very regular interval. This is also where it’s a good idea to start looking at the themes of what’s performing well and what isn’t and putting the “Battle Ship” method to use.
Review our “Creative Testing in Action” with our past client, DNBR.to to get some idea of how to iterate and continue testing.
This part is where the real creativity comes into play. The real insights into why customers choose to interact with your business really come out here and we’re frequently surprised by the results. This is the part of paid advertising that’s always interesting, exciting and keeps us in the game.
As always, this can be a complex process, so sign up for our newsletter or reach out to us and we can jump in to help you reach your goals.