Optimizing ads for ROAS will require healthy data practices and strong Custom Label usage. Here’s how to turn the average feed-based ad into a high performer.
Performance marketers will happily spend afternoons analyzing campaign results and making tweaks to campaign settings or demographics. However, these optimizations are only treating the final distribution of your ads. Those ads come straight from a product data feed, and, if that data is poorly constructed, campaign optimization is like putting a set of fake wings on a caterpillar.
The power of feed-based ads rests not only in the campaign settings, but the product data feed. Facebook, Google, Booking.com, Trivago—each of these rely on the data feed. When you want to optimize your ad images, copy, and even get them sent to a more relevant audience, it’s not just your campaigns that need to be optimized: you need to optimize your data feed.
Use custom labels
Perhaps most importantly, no attribute is growing in popularity quite like the custom label.
What are custom labels?
This attribute is used to segment products however their marketers see fit. You can use up to five custom labels in your feed at a time, labelled 0 through 4, with up to 1,000 unique values in each label group.
For some, marketing based on more detailed product types will be useful. For others, bidding decisions will be better made using ROAS or product margin data. Especially for larger brands, this is a step that should not be skipped, as custom labels not only allow you to monitor and track results more easily, they can also help find answers to more complex questions, like: do price ranges affect the performance?
Even if two products are very similar--even if they fall into the same Google Product category!--they could be very different in terms of how they perform and are best marketed.
0 ** | sale |
1 | price band |
2 | clothing and accessories |
3 | beauty |
4 | hero product/most popular |
Most or least popular label
You can only push so many products at once. When it’s not clear which products to push or where to throw your weight, knowing exactly which products are “most popular” can help. This allows you to focus on profitable products in campaign mode. Use it to focus on overall best sellers or top picks for certain segments.
“Loser” products, on the other hand, should be clearly marked, so you don’t waste resources on them. “Loser” products could be anything: those with low CPC, ROI, views. However, for products that are truly poor performers, consider removing them entirely.
Categories label
Many companies will opt to use custom labels to mimic product categories or to optimize those categories for campaign purposes. If you sell numerous types of products, consider using one label to denote groups like “accessories,” “apparel” or “beauty.”
Clearance label
Products on clearance are unique. For example, they may need to be pushed out fast to make room for new inventory. They also provide unique marketing opportunities, as adding “clearance” to any ad instantly changes the tone. Products labelled for clearance can be segmented from regularly-priced products and pushed to specific audiences more (or less) vigorously. Use this to quickly optimize special offer campaigns.
Price bracket label
How much is that click really worth? Adding price bracket information allows you to manage product groups by cost. Products that are expensive may be more difficult to sell. On the other hand, products that are cheap may be easier to market but have minimal returns.
Using this label allows you to suppress sales that are less desirable and focus on products that are in the sweet spot. If you have clear, defined audiences, matching different targets to different price ranges could be a practical no-brainer.
Seasonal label
You won’t be pushing snow shoes in the middle of summer. Mark seasonal items with custom labels to ensure you aren’t paying for ads at the wrong time. Plus, this will also make it easier to quickly manage those items that are season-appropriate and deserve a little boost.
Historic ROAS or ROI labels
Much like price bracket, historic ROAS information lets you automatically group products based on how they’ve already performed. Add existing ROAS information to better allocate funds and resources in your current campaigns.
High or Low Margin Labels
Not all clicks will make money. Label which products are particularly valuable and which are not. This will allow you to directly see and manage products according to the actual ROAS they provide. It can also assist in the planning process, as there will be a more clear picture of the margin levels and returns you can expect.
Examples: Possible products and their custom labels
__Product__ | __Custom Label__ | __Action__ |
Fancy Feline Cat Mug - Silver | Most popular | Increase bids |
DC Slip-On Shoes - Red | Least popular | Suppress bids or remove from feed |
Happy Sprinkle Sugar Mix | Clearance | Increase bids and review target audience |
Leather Pumps with Kitten Heel | High price bracket | Review target audience |
Santa Hat - XXXXXL | Winter | Change bids or remove from feed according to season |
Vegan Marshmallows, Silky Smooth - White | Low ROAS | Lower, monitor, or suppress bidding |
Luxury Wedding Tiara | High-margin | Increase bids |
Note that you may need to evaluate your current product data management solution in order to use custom labels, as not all modern solutions have the capability to offer custom label management to its fullest extent.
Exclude unprofitable products
You shouldn’t be exporting products that will damage your ROAS. Stepping back and removing them from your data feed exports is the single easiest way to ensure only ads that lead to profit end up being served.
Tracking data, like that from Google Analytics, should help you understand the real end-value of your ads. Use your metric of choice (ROAS, ROI, margins, conversions) to decide which products need to go. This will save you the hassle of dealing with unprofitable products on your export channels.
__Simple product exclusion__ | Exclude all products where: Margin < 15% | |
__Advanced product exclusion__ | Exclude all products where: Clicks > 100 AND Orders < 1. *Don’t spend more on clicks than you earn with product sales | Exclude all products where: Margin < 20% AND CPC > $2 *Don’t spend more on clicks than you earn in ROI |
Pay attention to your titles
Titles are often one of the few attributes a user will see and read, and optimizing them with the most relevant information leads directly to more clicks and more sales.
One agency tested 136 products over six weeks and found that making small changes to data feed titles paid huge dividends.
*“Since each product was optimized for a specific, highly relevant search query, we also took a look at the CTR uplift of the exact query match to the term we added. For example, if the term we added was “Adidas Slide,” we checked the CTR for queries that included exactly “Adidas Slide.” For these “built-in queries,” we saw a CTR increase of 88 percent!”* - Search Engine Land
The most competitive titles will be the result of careful planning and testing. That means you’ll want to consider adding one, two, or all of the below:
- Brand
- Search terms
- Product category
- Model number
For example:
__Unoptimized__ | __Optimized__ |
Mens Polo Shirt - Red | Lacoste Mens Polo Shirt, Slim Fit - Red |
Razer computer mouse - Black | Razer 2014 Ergonomic MMO Gaming Mouse - Black |
Search terms in particular should be leveraged when possible to drive returns. They already play a role in paid advertising, but they can also power up your feed. Find the exact keyword your audience is using and searching, and add it to your titles. This will help your audience better “click” with your product listings.
Using search terms is one of the most quantifiable steps you can take towards optimization.
Work at product level
There are plenty of reasons to make your product titles somewhat homogeneous. Keeping titles, descriptions, and images consistent can make managing tasks easier. However, it also completely glosses over the fact that every product is different. They’ll have different blockers and different possibilities. Some may need multiple images, while others would benefit from extra attributes.
In order to best express what makes each item unique and desirable, perform product data optimizations at the SKU level when possible, especially for valuable products. Treat your red celebrity-themed socks different from your costume hats--their audiences are different. Specifically, optimize titles uniquely and design images in accordance with their particular audience. Shoppers looking to buy a branded product will want to see that brand name quickly. Those searching for highly specific styles will want to see an exact model number they can reference, and basic items may benefit from powerful, SEO-optimized keywords. Plus, those audiences will have different expectations when it comes to imagery.
Though SKU level optimization isn’t practical for every last product, audiences for different products will search and respond in their own unique way.
Give your feed the mobile treatment
Adding mobile is about the greater Digital Transformation Imperative that ecommerce is facing today.
iProspect found that mobile paid search clicks continue to rise steadily, quarter after quarter; in fact, mobile devices accounted for nearly 62% of Google clicks in 2017. That’s up from 53% one year prior. And eMarketer also documented steady growth in mcommerce with mobile accounting for 58.9% of digital sales in 2017. This is why most marketers are already optimizing campaigns for mobile.
While feed-based marketing and shopping channels put increasing emphasis on their own interfaces’ mobile-friendliness, you can make sure your feed is mobile-friendly too:
- You can provide mobile deep links to ensure that shoppers end up on your mobile site or app, once they’ve clicked an ad--instead of being turned off by an unoptimized desktop page and exiting.
- Use only mobile-optimized sites. Desktop sites are functional on mobile, but far from ideal. Make numbers and buttons easily clickable. Simplify forms, navigation, and zooming. Most importantly, make purchasing easy.
- Keep titles and descriptions in the character limit. That means checking the limits for your most important channels. Amazon titles will show less than 70 characters on mobile and less 200 characters for descriptions. When in doubt, preview it for yourself.
- Optimize images. Google Shopping and Amazon in particular outline numerous sizes and shapes for different mobile settings. Facebook, on the other hand, prefers to use the same images.
Optimize your best performers
High number of clicks, low CPC, and good margins all make a product great. Positive results like these mean that the product is well-targeted, the ads are well-crafted, and the users are willing to buy. With already positive results, feed optimization can push that ad to a wider audience with more predictable results.
Some products will simply perform better; be sure your most valuable products are pulling in their absolute maximum returns possible.
Start by locating your top performers. This can be done manually by checking your individual export channels and reviewing the results; however, this can take time and resources. The process is better done through a feed management tool that can pull all performance data together, highlighting which products are working best overall. Then, for further enhancement, you could go deeper into enhancements for individual channels.
When you get into the nitty gritty and optimize each for your different export channels, keep in mind the 5 Golden Rules of a Quality Data Feed.
Make those feed optimizations happen
With your realistic expectations and a clear goal set, it’s time to move forward and start optimizing for ROAS. And once you’ve done all of the above, review your results and keep optimizing and measuring.