Dynamic Remarketing is one of those things that seem slightly more complicated than it actually is. But at the same time there are definitely things that you should know before just start spending money on dynamic remarketing.
In this post we’ll get into some of the big picture things like whether it even makes sense to run remarketing for eCommerce stores to the nitty gritty with how many audiences you should create and what ad formats to use.
In essence Dynamic Remarketing is the following:
Traditional remarketing means that someone visits your ecommerce store and then you target then with ads afterwards.
Dynamic Remarketing is the same thing, just with dynamic ad content.
The content in your ads will change depending on what products the user was exposed to on your site.
Hence – dynamic remarketing.
The way Google knows what products to show to the user is because you insert a piece of code on your website that changes depending on what page it’s on:
(not full code – see complete implementation guide here)
And that’s it 🙂
Dynamic Ads Are Best for Product Page Visitors and Shopping Cart Abandoners
There is no doubt that where you’ll find the highest ROI is from visitors that have been to product pages or shopping cart abandoners.
It’s not coincidental that these are the two places where you narrow down the products you’re advertising for a lot and that users are a lot further in the buying cycle when they visit these pages than the other pages (search, category, home) on your site.
Another thing you should keep in mind is that categories typically contain a lot of products. When Google is then trying to showcase a couple of products in the dynamic ads then Google will choose randomly from the (many) products. These ads are typically not that effective.
For this reason I often recommend using static banners for all other pages than product and shopping cart audience lists.
Dynamic Remarketing Ads: How to “Design” Great-looking, Effective Banners
Dynamic Remarketing banners are among some of the ugliest banners known to man.
The font selection is horrible.
The logos are almost always squished or too small to see.
The buttons are outright ugly.
But they work.
So if you were like me and discarding dynamic remarketing because of the way the banners look, then I recommend that you reconsider.
I typically advocate starting out with two different ad sets:
1) Your favorite layout
2) Another layout where you give Google the freedom to test the various layouts to find the best performing one in each size.
I’ve seen that Google often perform pretty well when it comes to testing out these banners. Mainly just because I think it’s cumbersome to test image ads in the first place. Add Dynamic Product Ads to it and it gets much more complicated.
No matter what you think of Google’s testing methodology, then it’s better than no testing at all.
Below I’ll go through some tips to make the banners look decent.
1) Choosing Your Own Layout for Dynamic Remarketing Ads
There are a number of layouts to choose from.
Which one you should choose is really up to you.
I typically recommend using a layout that pulls in ~3-4 products in the same ad. It’s rare that users only visit a single product on your website, so this will enable you to show more products, and related products, to the user.
Colors: Try to go with your brand colors as much as possible.
Button-Text: Avoid just going for “BUY NOW” like every marketer known to man like to use. Trying something less invasive like Browse, Read More, Free Shipping or similar wording instead of just buy now often results in better click through rates.
What you put on the buttons in your remarketing ads will not make or break your performance. Truuust me.
If you do this, then you’re able to create passable designs for Dynamic Remarketing.
2) Choosing the Google-Auto Layout
The other option is to let Google test the various layouts and run with the layout that has the highest performance.
I’ve seen the best results with this. Google routinely change out the banners and test new versions that just come out. It can be tough to know exactly when new layouts are launched, so it’s an easy fix.
Just don’t skip adding your brand colors and the other items listed above. You still don’t want an Internet Explorer blue button on your banner 🙂
Designing Your Own Banners
I’ve personally not done it yet, but it’s on our to-do list in 2018 to test out our own banners versus Google’s.
I think that once you’ve proven that you can generate a profit using dynamic remarketing within AdWords, then you should test designing your own banners.
It’s not as easy to design these banners as you normally would. You can’t just hire any design off Fiverr and give them some copy.
The designer needs to know how to insert the proper tags in the proper places to pull information from your Google Merchant Center.
#PPCCHAT Is Undecided
A quick poll on Twitter is almost completely undecided between the three options:
The Remarketing Lists You Should Create For Maximum ROAS
This is where the advice you see across the web shifts, a lot.
If you’re reading advice from someone running campaigns for massive eCommerce stores you’ll see them recommend using dozens of very granular segments.
But if you’re a mid-sized eCommerce store you’ll look at those recommendations and realize that even though it sounds smart to segment very granularly, then you’ll most likely not have enough users for each segment.
The standard setup that I recommend the most is setting up the following lists:
- All visitors: 7, 30, 90, 540 days
- Product visitors: 7, 30, 90, 540 days
- Category visitors: 7, 30, 90, 540 days
- Shopping Cart: 7, 30, 90, 540 days
If you have other sections of your store with a lot of traffic (FAQ, blog, etc.) then you might want to use these for remarketing as well. However, I wouldn’t advise that you add them to begin with and I wouldn’t use them for dynamic remarketing.
If your average buying cycle is longer than 7 days, then you can change the 7 day list to a 14 day list.
If you’ve build your audience lists in Google Analytics I recommend adding another layer to any of these:
- All visitors +1 PV: 7, 30, 90, 540 days
- Product visitors +1 PV: 7, 30, 90, 540 days,
- Category visitors +1 PV: 7, 30, 90, 540 days
PV = pageview.
The extra layer should be anyone that have visited more than one page on your site. The goal is to exclude the people who merely visited one page on your site and bounced. You can also choose users who spent less than 60 seconds on the first page. Whatever fits your exact formula.
The Thinking Behind 7, 30 and 90-day Lists
The fresher the user is, the more likely it is that they will make a purchase.
The main idea behind breaking up your remarketing lists in separate days is to bid higher for your 7-day lists than your 90-day lists.
Your bid for the 7-day list might be $1, while the 30-day will be $.60 and the 90-day will be $.4. Or whatever makes sense according to your own ROAS performance.
The 90-day list is really just for catching any potential visitors that might buy something, but it’s rare this audience provide any meaningful sales volume. Also remember that if someone comes back to the site after being on the 90-day list and then leave again, then they’re moved back to your 7-day list. Of course, only if they actually click a banner while being on the 7-day list will the sale be accredited to that list. But I’m ranting 🙂
Further down I’ll tackle sequenced remarketing. The different 7, 30 and 90 day segments will be used for this too.
But 540 days, Andrew? I don’t even remember what I had for breakfast!
The 540-day list is not intended to be used on a daily basis, and definitely not in dynamic remarketing. The idea is simply to use this list for holiday sales or other times where you want to get out to as many people as possible.
If you know XYZ person has actively searched for beauty products and visited your site in the last 1+ year, then chances that they will be interested in seeing what you have to offer during a promotion much higher than the general audience. You can use this list for Remarketing Lists for Search Ads, but also for general banner advertising. Just don’t think that it’ll be as effective as traditional remarketing.
Honestly, I’d love to be able to set a small delay in when the remarketing ads kick in. I’ve always had a love/hate relationship with remarketing ads that kick in immediately after you leave a website.
On one hand, it helps reinforce the name/brand of the store you just visited.
On the other hand,then there is no way that I will leave a site, see an ad 30 seconds later and then finish my purchase. Give me an hour at least 🙂
Let Your Campaign Run for a Minimum Of 30-60 Days
One of the observations I’ve had around Dynamic Remarketing in Google AdWords is that they get better over time. Especially if you enable Google to choose the banner layouts themselves.
They’ll test the various banners and slowly move into sticking with the better performing banners.
I highly recommend that you let your remarketing run for that reason. If you’re concerned about the unprofitable spend, then try to decrease your bid for your All Visitors and Category visitors lists.
The biggest bang for your buck will be from your product page visitors and shopping cart abandoners anyways.
The Small Stuff: Settings
Frequency capping is one of those things that no one really have the final answer to.
I typically advise setting it to 10-15 impressions per user, per day, per campaign. That way you know that it’ll not run amok, but still give more than enough room for the campaign to reach users.
Remember to set your location settings to People Located In instead of People Located In or Interested In
If you set it to interested in then you’ll also target people who visited your site, but are sitting outside the country you’re targeting. Typically, this is a waste of money.
Conservative or Aggressive Targeting
This is an interesting little nugget that Google has slipped into the settings as a “default”.
It’s also known as the Google Display Optimizer:
It basically means that Google reviews the website (or placement) that your ads are showing on, and then it will try to find similar websites to show your ads on. This is not remarketing. By not disabling this setting, you’re moving into traditional display advertising.
I recommend that you tick off both Conservative and Aggressive Targeting. If you want to get into display advertising generally, then you should pursue this, but don’t do it with your dynamic remarketing campaign. The ads are not suited for success generally on the display network.
Exclude Mobile Apps
You can quickly spend 2-3-5x of what you should if you include targeting for mobile apps.
In the old days, you’d have to exclude the domain or website called adsenseformobileapps.com to exclude your ads from showing on mobile apps.
Today, we’re more sophisticated.
When you’re setting up your campaigns you can easily click to exclude Apps from AdWords display advertising / remarketing:
I advise that you:
- Exclude Tablet Apps
- Exclude Mobile Apps (and consider excluding all mobile)
I haven’t had any success with targeting mobile devices for display advertising.
If you really want to target mobile apps or mobile users in general (there are definitely cases where it makes sense!), then create a dedicated campaign just for mobile. It’ll make it easier to track and make the proper changes in order to make it profitable.
This approach creates more work than you need, but way too often I see ecommerce advertisers “forget” about devices in their campaigns when it’s really the mobile part of a campaign that’s not performing as well as it should:
Questions from the Audience
Should I run Facebook Dynamic Product Ads Simultaneously?
Dynamic Remarketing ads with Google cover only websites that are using Google AdSense (or said in plain English: Only websites that have opted into Google’s display network in order to run ads from Google advertisers).
In order to access advertising slots on Facebook, then you need to advertise directly with Facebook (or third parties that have access).
So in order to get the best possible reach, then using Facebook Dynamic Product Ads alongside your Google Dynamic Remarketing is the way to go.
Should I Run Criteo/AdRoll/Other Product Ad Remarketing Simultaneously?
My short, and standard, answer is no. Google has a very big part of the display advertising slots covered, so you’re in most cases covered with just using Google and Facebook.
Yes, other 3rd party remarketing tools do have access to some advertising slots that Google and Facebook don’t have access to, but it’s not much.
The way to think of Facebook, Google and 3rd parties is the following:
Facebook offer advertising slots across Facebook, Instagram and a few partner sites/apps.
Google offer advertising slots across their Google Display Network.
3rd parties usually offer advertising slots on the two above-mentioned places as well as some other markets they control themselves or are renting.
So no. There isn’t a lot of advantages with going further than Google and Facebook when it comes to remarketing. As always, this is a standard advice – if you’re already maxing Facebook and Google Remarketing, it’s always worth a try to find new channels for growth.
Are Promotions/Coupons Necessary for Remarketing Success?
I urge most advertisers not to give away specific coupon codes in their remarketing.
I know it might be effective to begin with, but you train your visitors to leave your site and wait to be remarketed to. Especially your repeat customers, the ones you’re supposed to make the most profit from, will get to know this trick and use it all. The. Time.
There are entire communities in Facebook groups where all they do is share what pages you should leave at what time and for how long to wait until you receive a discount code.
Would I have received these sales anyways?
When running remarketing campaigns there are always that gnawing question of whether you would have been able to get the sale without using remarketing.
I generally think that you can’t make it into such a binary, 1-to-1, decision.
When I hired the first person for Savvy and needed to buy new equipment I couldn’t use my regular stores. Filip is located in Denmark, so I needed to use Danish stores. The thing is – I haven’t bought anything online in Denmark in 8 years. So I couldn’t really go to my regular stores.
I do Search a lot in Denmark though as we have a significant amount of clients there. So I had some knowledge of the different stores that were available.
However, I started like anyone else does and went to some of the sites I know. Elgiganten.dk was one of them. Power.dk was another. I didn’t find what I was looking for, so I searched for random keywords like computer monitor or lg monitor to see what other sites were there.
I clicked on Shopping ads even though I knew I didn’t want the product that was displayed. I thought, well if they have this monitor, then they should have others too (this surprises a lot of store owners – many users don’t purchase the product shown in the Shopping ads).
As I gradually narrowed down my search, I knew exactly what monitor I wanted so I searched for that specific monitor and chose the cheapest store that was also convenient
Convenient for me equals:
- Has a large selection of other items so I only have to buy at one place
- Enables me to just buy from their again without much hassle
- Allows for quick/easy shipping
- Good reputation – I don’t want to have issues when I need customer service
After that I then started to look for a headset. I googled a bit, but then remembered I knew headsetshop.dk from yeeeaaaars ago. I went directly to the site and bought a headset. How the heck is headsetshop.dk going to credit me to any marketing?
- In my specific case, there was no need for a remarketing display ad, so yes – the sale still happened
- When buying a headset, I will not rule out that I had been nudged by remarketing ads in the past when visiting headsetshop.dk
When it comes to whether it’s worth running Dynamic Remarketing then there are two things you should do:
- Implement the Position-Based attribution modeling in AdWords
- Review your days to purchase report in Analytics to see how long it generally takes a customer to go from visit to purchase
Generally, you’ll be able to get a lot out of the attribution model comparison tool in Google Analytics (and yes, to all of you who say that “attribution is flawed”, yes we all agree – but it’s what we have currently and it’s better than nothing). Compare the revenue/sales attributed to Dynamic Remarketing between the Last Interaction and position-based. You can even use linear or a custom attribution model to do it.
Is Remarketing for Everyone? I See Everyone Saying It’s Amazing, But It’s Not Working for Me
The more expensive the product, the longer a purchasing decision takes. This directly impacts how impactful remarketing will be. The longer a purchasing decision takes, the longer you’ll have to show remarketing ads.
I also see that dynamic remarketing tends to take a bit of time to gain momentum. You can’t just look at your campaign in its first month and say: “Oh well, I tried it”.
Just like with any campaign, then the key to success is finding the variables that aren’t working for you. That’s the only way to grow your AdWords account instead of remain stale.
Sequenced, or Paused Dynamic Remarketing
I’ve always thought starting an ad the second someone leaves your site is very aggressive.
But that’s not the idea of what we’re talking about here.
Sequenced remarketing means that you show your users one ad the first 7 days, then another ad for 7 days, then another ad for 7 days, etc.
I generally think that showing the same boring Dynamic Product Ad nonstop for 90 days isn’t very efficient.
Instead, I urge you to do this:
For the first 7-14 days, push your Dynamic Product Ad:
After 7-14 days, try a general image ad that you have somewhat targeted to what the user looked at.
So if your visitor looked at headphones, then create a banner ad that focuses on your top headphones.
The entire idea with using a banner that you design yourself is that you can push other marketing messages than “just” the image of your product.
So ensure that you follow proper banner design guidelines:
- Don’t overcrowd it
- Focus on what’s most important
- Easy to read text
- One headline, and 1-2 supporting points
This means you have to build some custom remarketing lists (again, sounds harder than it is).
Should You Exclude Converted Users?
One of the common best practices is to exclude users that have recently converted from your remarketing efforts in display.
On the flip side though, I’ve audited accounts where there have been a positive ROAS from the converted users segment in Display Remarketing. Really?
Personally, I don’t get it and I haven’t seen enough of these to advise anyone to not exclude converted users. I say let the email team go after people who have recently converted.
Note, if a converted user goes out and searches for the product again, then in most cases they’re fair game and I don’t advise excluding them for Search ads.
What ROI should you go after?
When you talk about Dynamic Remarketing, you’d be amiss if you didn’t discuss the attribution aspect and whether you should go after another ROI for your remarketing campaigns seeing that you’ve “already paid” for the users you’re marketing to.
With that, I urge you to consider the following:
You’re almost always advertising for people you’ve advertised to before. Think of how many times you’ve seen the same AT&T ad? Or driven past a Microsoft billboard in Seattle?
Thinking that remarketing is the only advertising format where you’re actively advertising repeatedly for the same user is ludicrous.
Where you have to start reviewing the remarketing ROI is when you’re specifically working on funnels.
Let’s take a simple case.
Let’s say that you’re running Facebook Ads where you don’t care about being profitable because you’ll get the profits with your remarketing campaigns. In such a case, it’s worth including your Facebook ad costs to your remarketing costs and ROI calculation.
Going back to the question, then I personally think you should aim a bit higher than your regular ROI target. So if you normally target 500% ROAS, then upping it to 700% might be a good move. But essentially, you should have an overview of all your marketing and more or less know whether it’s making sense or not.
Should I Count View-Through Conversions?
When I’m asked this question it reminds me of the discussion of attribution models.
It’s not as much about whether view-through conversions should be used or not. It’s not a binary question like that. It’s more of – when should you use view-through conversions.
I would never use view-through conversions to try to make up for the fact that a campaign wasn’t making money.
But if a remarketing campaign is not being profitable, but it’s seeing a lot of view-through conversions, then I would definitely keep it alive for a longer time to try to find out why it’s not converting “directly”.
On the same note. No real conversions and no view-through conversions is definitely a strong sign that you’re either doing something wrong with your dynamic remarketing (95%) or your store/product/market is not suited for dynamic remarketing (5%).
- Dynamic Remarketing increases in importance the longer your buying cycle is
- The majority of your success will come from product page visitors and shopping cart abandoners
- A quick setup uses the automatic layout testing from Google, which has great success
- Mix in regular banner ads for your remarketing to be more successful
- Segment your lists based on how long it’s been since they’ve visited your site
- Exclude mobile apps – please just do it
All in all, Dynamic Remarketing in AdWords is a campaign-type that all ecommerce advertisers should use. It’s a great way to automatically create highly-specific ads with product information in them that targets users that have seen these products before.