By Amy-Louise Adlington

As with all things PPC, we at Merkle | Periscopix have a best practice for structuring Shopping campaigns. But what if that’s not working for your client? What if sometimes you need something different to get top results?  We used custom labels to create an alternative campaign structure for our client Simply Home Entertainment, and the result was an 88% improvement to ROAS.

Back to blog home

The Standard Old Structure

When we first set up Simply Home Entertainment’s Shopping campaign, we used one of our trialled and tested structures. A campaign split into multiple ad groups via “product type”. (For a full run down of some of our top Shopping campaign structures see this article.) However, this didn’t showcase the bestselling points of their stock, and performance showed this. We had three main problems: 

  • We were accruing a lot of traffic to site. This meant big spend and it was challenging to keep to budget.
  • Even more granular products groups such as “documentary DVDs” or “rock CDs” were just too broad as products within them were extremely varied; encompassing big blockbuster titles as well as specialist and niche ones. This made optimisation nearly impossible as there was no pattern to product group behaviour, and optimising at product level when you have 1000’s of products just isn’t a viable option.
  • Though we were getting a lot of traffic, pitched against big brand competitors we just weren’t seeing equivalent conversion volume. Consequently, ROAS was suffering.
Table showing the old campaign structure

The Shiny New Structure

We decided that the standard structure just wasn’t working, and came up with a plan to tailor our Shopping structure to the unique selling points of Simply Home Entertainment’s stock. We brainstormed with the client, and came up with the top four USPs of their stock. These were:

  • “Simply Media” titles.
  • Titles featured in the monthly catalogue.
  • Collectors box sets.
  • Other miscellaneous niche titles.

We then created 4 corresponding custom labels, which would then become our 4 main ad groups (with an additional catch all ad group set at a low bid to complete the new campaign). Each of these ad groups could then be split into more granular product groups by product type as before.

Table showing the new campaign structure

Remarkable Results

  • Spend for this campaign was initially reduced by 49%. We reinvested our saved budget to further boost performance. This included maximising impression share for our 4 main ad groups.
  • Conversion rate improved by 21%.
  • ROAS improved by 88%!
  • As the ad groups now contained products with the same USPs we began to see patterns to performance, which made optimisation a lot easier for us. Time was saved, meaning we got to spend more on strategy.
Graph comparing weekly ROAS of the two campaigns over separate 8 week periods

Lessons Learnt

While tried and tested set ups are a good place to start, e-comm clients with more niche offerings can really benefit from some out-of-the-box thinking and a customised structure. We’ve continued to see great performance from our new structure, which has allowed us to try some new strategies. More consistent conversion volumes (as well as consistently running safely within budget) has enabled us to try a target ROAS bid strategy on the Shopping campaign. This has resulted in increased traffic to site and the campaign generating more conversions than ever! We’re now looking to boost our generic coverage via DSA ads, carrying this structure over to a DSA Feed campaign – hopefully we’ll see similar success!

Share this article