Shopping ads that show the right products, and Microsoft ads that cost less
In Shopping, the feed is the campaign. There are no keywords to write. Google decides which searches your products appear for based almost entirely on the titles, attributes and images you send it, which means a neglected feed quietly limits everything downstream no matter how the bidding is managed.
Microsoft Ads is the other half of this page and the more commonly skipped one. It is smaller, the clicks are meaningfully cheaper, and for B2B and older demographics it converts well. What it is not is a copy paste of your Google account, which is why most imported Microsoft accounts underperform and then get abandoned.

Start in Merchant Center, not the ad account
Before a single bid is set we look at how many products are actually eligible to serve. Disapprovals, missing attributes, image problems and price mismatches routinely take a fifth of a catalog out of the auction without anyone noticing. Fixing that is free reach.
Titles are the keywords you are allowed to write
Shopping matches on the feed, and the title carries most of the weight. Brand, product type, key attribute, size or model, in the order a person would search it. Rewriting titles across a catalog is unglamorous work that regularly moves impressions more than any bidding change.
Split by margin so bidding can be honest
When every product sits in one campaign, the algorithm optimizes toward whatever converts most easily, which is often your thinnest margin item. Splitting by margin and price band lets you bid what each group is actually worth and stops the cheap products consuming the budget.
Performance Max needs the same guardrails here
PMax carries most Shopping inventory now, and it will happily spend on brand terms you were already winning. We use brand exclusions, split asset groups by product group, and keep enough visibility to know what is being bought.
Microsoft is a rebuild, not an import
The import button exists and it is where most Microsoft accounts go to die. Match types behave differently, the competition is different, and audience targeting is stronger. We import as a starting point and then rebuild structure, negatives and bids for that auction specifically.
Judged on return, product by product
Total revenue hides everything. The report shows return by product group, which products are spending without converting, and what we are changing. If a category cannot make the maths work, we would rather cut it than average it into a healthier looking total.







