Our innovative profit-based approach to bidding drove a 58% cost reduction for Cheekywipes significantly improving the profitability of their paid search channel.
Upon an initial audit of the account, we saw opportunities for optimisation in four key areas.
- Overly granular legacy structure: Cheeky Wipes had previously adopted a granular structure within Google Ads, running 21 campaigns in total: 18 search and 3 shopping. Unnecessary traffic segmentation can often negatively impact the performance of campaigns due to the reduced learning of the algorithms.
- Insufficient data to bid towards profitability. Products were grouped by margin, as we did not have sufficient data to bid towards profitability. We therefore needed to import basket-level profit data into Google Ads, which we could use as the basis of our bids.
- Inefficient bidding strategy: Margin-based bidding had not yet been implemented on the shopping campaigns, so we were not optimising towards higher margin products. Manual bidding was being used, which does not take into account all variables, and does not depend upon the efficiency of the algorithm.
- Lack of optimisations: The shopping product feed required more optimisations to further improve its performance.
- No audience-based bidding: Audiences had not yet been imported into Google Ads, so bids were not being adjusted according to new or returning customers.
We took a multi-pronged approach to improving Cheekywipes paid channel, firstly focusing on structural consolidation, implementing profit-based bidding, optimising their product feed and moving to a CSS as well as getting the most out of their first-party data with some clever audience segmentation.
With Google giving advertisers access to their query-level bidding technology (Smart Bidding), it is key that advertisers give Google enough data for the technology to work effectively.
To do so, Google requires at least 50 monthly conversions per campaign. When we acquired the management of Cheeky Wipes’ paid media account, they had a very granular structure with a large number of campaigns and one keyword per ad group.
We reviewed the keyword historical conversion volume and then regrouped their keywords, ensuring that each campaign would reach 50 conversions per month.
We proceeded to simplify the account to a more consolidated structure, and decreased the number of campaigns to 5 split by keyword intent of brand, non-brand and competitors.
Then, we consolidated the ad groups from one keyword per ad group (SKAG) to having several themed keywords per ad group (MKAG). This focused on driving 3000 weekly impressions per ad group, so that the RSA (Responsive Search Ads) would have enough data to optimise effectively.
To ensure that searchers would receive highly relevant ads, we used DKI (Dynamic Keyword Insertion), which inserts the most relevant keyword into the headline of the ad and therefore delivered highly personalised ad copy to users.
Profit based bidding
The challenge around bidding towards profitability
Cheeky Wipes has a large catalog of products with different margins and different shipping costs, which originally made it challenging to approach optimising for profit.
To achieve the best possible results with Google’s Smart Bidding, we needed to give Google profitability data for its algorithm to start optimising towards.
When we started working with Cheeky Wipes, they had their products grouped together by margin. Whilst this enables you to set ROAS targets based on the margin of products within that grouping, there were three major issues that were overlooked and had to be addressed:
Firstly, 80% of Cheeky Wipes’ orders contain more than one item, so at basket level you would have a blended margin of the products, not purely the margin of the product they clicked on.
Secondly, we identified that within a large number of orders, people had not purchased the item that they clicked on. This meant that the item that was bought would likely have had a vastly different margin.
Finally, it did not take into account the shipping cost, tax or payment processing fees that vary from product to product.
The solution: profit based bidding
To bid towards profitability, we required net profit data in Google Ads. We needed a solution that would help us:
- Look at the basket holistically, taking into consideration the margins of different products
- Take into consideration order costs, such as shipping and VAT
- Take into consideration overhead costs, such as staff and agency costs
Gathering this data would give us a more complete picture of profitability inside Google Ads, and allow us to optimise towards true net profit, as opposed to optimising towards ROAS.
To achieve this, we partnered with a specialist data platform called ProfitMetrics that would allow us to connect the dots and link a thorough set of data together.
We also worked alongside Cheeky Wipes’ development team to implement the code of the website. This gave us full visibility of which products were in the checkout, how much VAT was charged and what the shipping cost was.
The second step was feeding the margin data for the products into ProfitMetrics. We worked with the commercial and operational teams at Cheeky Wipes to collate a spreadsheet with all of the product level cost of sale data, as well as the costs to ship different products at different weights.
Ultimately, we added our agency fee and Cheeky Wipes’ overhead costs to the ProfitMetrics platform.
We had finally gathered all of the data that we needed into one system, allowing us to accurately determine profitability. Inside ProfitMetrics we could see order level data, with full product margin and relevant costs including taxes, shipping and overheads.
Better still, we were able to set up a conversion action within Google Ads allowing us to pass net profit data into the platform as conversion value.
Optimising towards maximum profit
Once we had the data within both ProfitMetrics and Google Ads, we were able to do three things to help us optimise towards maximum profitability.
- Firstly, we fed the data in from ProfitMetrics as revenue into a conversion action and we moved the bidding strategy from tROAS to maximise conversion value. Google had all of the data that it needed to truly maximise for profit, while taking into consideration all of the relevant costs such as delivery, VAT and overheads.
- The data we gathered gave us the opportunity to segment the products by business objective, as opposed to only segmenting by margin or product category. We could overlay segments such as produce level price competitiveness or stock levels, allowing us to optimise towards these secondary objectives.
- We built a model using the data below to work out where diminishing returns would be hit and where the peak profitability would be achieved.
Using Google’s budget simulator we were able to model out when our gross profit would be. This is defined as our (gross profit = revenue – cost of goods sold – shopping – transaction fees).
This was determined as we set our primary conversion goal to Gross Profit, and we had been feeding the conversion data in from profit metrics to allow Google to learn.
We then determined the net profit by subtracting the media spend (net profit = gross profit – media spend).
We graphed this data out below to show where the budget was required to hit the maximum amount of net profit from the channel.
Following Google being fined by the European Competition Authority, they introduced CSS (Comparison Shopping Service), which effectively allows competitors to enter the Google Shopping results at up to a 20% discount in CPCs.
Cheeky Wipes were still using Google as their CSS, which we saw as a quick win, and one that would enable significant cost savings.
We implemented a third-party CSS called Shoptimised for a nominal fixed monthly fee, as opposed to the 20% of the click price that Google Merchant Center CSS charges.
With Google’s PMAX becoming more black box, the data that you put in, especially the product feed, is the lever that you can pull to improve performance.
As a starting point, we ran a feed audit with Shoptimised and gathered general suggestions, such as adding colours and missing attributes to the products or adjusting the titles.
Secondly, we created a priority list. We exported the shopping report from Google to identify which products were driving the most sales. We then created a bucket of the products that generated 80% of the sales, which were our top-performing products, and then a second bucket of those remaining.
The top performers would then be manually optimised, and the second bucket containing the remaining products would be primarily optimised through the use of rules that would find and replace terms.
We worked through the original suggestions rigorously, starting with the product titles. Having consulted with CheekyWipes, we determined that we wanted to focus heavily on brand protection as opposed to trying to find more new users with generic terms.
Based on this, we also decided to move the brand name to the front of all of the product titles. This helps Google understand when to show the products, as it favours showing the ads for terms that are towards the front of the product titles.
Furthermore, as Cheeky Wipes product catalogue doesn’t fit into a conventional title structure, such as fashion or electronics, we decided to create our own title structure as follows to give Google the best information in the right order: brand, product name, size, patent, colour.
We then used the advanced rules within Shoptimised to complete the missing information that was contained in the titles but not in the attributes. For example, if the title of the product contained the colour “blue”, but the colour attribute was empty, it would insert it into the colour attribute.
As part of our initial audit, we identified that many of their products were missing MPNs and GTINs, that Google use to help determine when to show the product. We worked with Cheeky Wipes to add these into the feed.
We also reviewed their Google Product Categories. These are often overlooked in the feed but, from our experience, play an important role in when products are served. Once again, we used the advanced tools within Shoptimised for suggestions, and then manually reviewed them to ensure products were in the most relevant categories.
Finally, we optimised the product images manually by reviewing the different images available for each product, and selecting the most appropriate.
Finally, we optimised our customer acquisition with audiences that automatically update. Cheeky Wipes started using Klaviyo, a marketing automation platform, to monitor these audiences, which we imported into Google Ads.
This integrated data and therefore gave the algorithm more signals to adjust bids and maximise the limited ad spend, such as down weighting bids on recent purchasers who, due to the reusable nature of the products, were unlikely to convert again soon. In comparison, we increased bids for email subscribers, and other audiences more likely to convert.
Not only did these optimisations improve the efficiency towards audiences more likely to convert, but also reduced cost, as we were not unnecessarily spending on audiences that had recently purchased.