Previously Paid Search activity focused primarily on online performance, meaning the influence of Paid Search on offline activity was not visible within our campaign outputs. Merkle and B&Q wanted to understand the true halo effect of digital activity and how it can influence offline as well as online performance. Ultimately, our objective was to:
By validating the true value of digital, the business hoped to justify further digital investment.
We approached this challenge by creating custom omni-channel formula columns in SA360. Due to data limitations which prevented us from accurately ingesting in-store revenue into our SA360 platform, we had to create a bespoke solution. Utilising B&Q internal data, we assigned a static value to store visits, allowing us to utilise formula columns in SA360 to create a proxy in-store revenue value. We could then amalgamate this value with our online metrics to create an omni-channel view of revenue. By integrating this new omni-channel revenue metric into an Auction Time Bidding Target ROAS bid strategy - we could effectively bid to combined online and offline revenue figures. This meant we could prove that paid search activations could uplift total revenue.
By bidding to a combined online and offline data source, we achieved a significant increase in total ROAS by 32%, total sales by 32% and total revenue by 19%. Crucially it helped us to start a conversation around the full value of digital – and how online-only measurements don’t account for total campaign value. It helped to highlight a key data ingestion issue across the B&Q brand and helped pave the way for more holistic campaign measurement and evaluation going forwards. In the short term we were able to obtain additional B&Q investment for Search Local campaign formats - further bridging the online to offline user journey. In the longer term, we have started to scope out full store sales integration with our online platforms – meaning that for future campaigns we should start to be able to pull through dynamic offline revenue figures (rather than the static store visit value figure used within this test.