What was the challenge?
We were tasked to optimise paid social activity for Debenham’s to increase overall ROAS for their womenswear department. Running ads across Facebook & Instagram, the client was achieving an average ROAS of 2.36 in-house, which was well below their target ROAS of 4.66. We ran always-on activity & campaigns during peak periods such as Black Friday. Below are some example creatives.
What was our solution?
We began with a full review on more than £650k of ad spend, identifying key issues and opportunities to optimise activity. There were some quick wins such as ad sets with high average frequency, which in some cases were as high as 14, so we capped frequency to increase reach and in turn, ROI.
Strategy rethink!
Debenhams had a strong partnership working closely with Facebook and applying their recommended DABA (Dynamic Ads Broad Audience) approach. This meant ads had little targeting other than age range & UK location but leaves the optimisation all up to Facebook’s AI. This can be good practice, however, it does mean the advertiser doesn’t have visibility on the learnings and takes away all the control. We balanced this by adding lookalike audiences of purchasers, testing from a 1% lookalike (most similar) up to 5%, which is now the best-performing audience and outperforming Facebook’s DABA approach.
Making the ads work harder..
Debenhams regularly ran new offers on product lines such as 20% off boots or 10% off jackets, however, as each new offer arrived, this meant updating the adverts too. Facebook’s algorithm works best once it achieves 50 conversions to get out of their learning phase but with every updated advert, it kick started the learning phase again. This approach meant campaigns were never working at an optimised state and capitalising on the previous learnings.
To counter this, we restructured the account, separating campaigns for remarketing & prospecting. For the prospecting ads, we implemented a dynamic feed allowing us to use a Google doc to amend offers without the need to create new adverts, in turn, this meant we didn’t restart Facebook’s learning! Not only did this make the campaigns work smarter, but it made campaign management more streamlined. It meant meeting deadlines or making last minute changes could be much more easily achieved.
Audience optimisation
The conversion-focused prospecting activity had 2 different objectives; website conversions and catalogue conversions. The campaigns were set up to achieve the lowest CPA (lowest cost strategy) which is often logical, however, while this may achieve the lowest CPA this doesn’t always translate to the highest ROAS. Debenhams often saw low average basket values of around £35, which meant that despite generating low CPA’s, ROI was not as strong due to the low ABV. Facebook no longer allows us to target users via higher household income who may contribute to a larger average basket value and a higher ROAS. However, Facebook does allow us to use a value optimisation bid strategy, which targets users that are likely to spend more and which may encompass some of this audience without directly targeting them. This strategy has performed incredibly well each week helping to increase ROAS and is one of the top-performing campaigns WoW.
Product sets
Using a 3rd party tool, we identified “hidden gems” or products which individually performed really well but not necessarily spending large amounts. We also identified high spending products that had below-average ROI and either removed or adjusted the exposure. Pushing different product sets helped to optimise activity further.
Results
In 2 months, we more than doubled the inhouse ROAS to more than 5.00 and exceeded the target ROAS of 4.66. We went on to achieve the best ever day in their history with a 5.28 ROAS.
Not only did we get excellent results in the competitive retail sector but having the opportunity to work with Debenhams (one of Britain’s most historic brands established 1773) while they transitioned to digital under the new ownership of Boohoo was an exciting prospect which presented its own unique challenges.