Properly Optimize Your Campaign with Your New Higher Quality Conversion Data

Now that we're pushing high quality conversion data to Google Ads for MQLs and/or sales using WhatConverts, it's time to make campaign adjustments based on past performance to drive more MQLs and sales.

I'm going to assume for this tutorial that you started a new campaign when you updated the conversion tracking.

When it comes to bid strategies, you can utilize Maximize Conversions now without worrying about that negative feedback loop that was generating low quality leads. You can also start leveraging other smart bid strategies that utilize conversion value like Maximize Conversion Value and Target ROAS bid strategies.

There are other manual adjustments you can do to improve the campaigns performance as well, I'll show you those now.

First, I want you to look at the last 30 days and make note of the following:

Conv. value / Cost (This is the ROAS)
Cost / Conv.

Now set the campaign back to 'all time' and make note of the following:

Conv. value / Cost (This is the ROAS)
Cost / Conv.

Now that we're still looking at all time data, let's proceed.

Keywords

First I'll look at the keywords on the campaign level, not ad group level. This shows us all the keywords in the campaign regardless what ad group they're in.

I'll then filter out any keywords that have generated a conversion and then sort by 'cost'. If any keywords have higher cost than what campaign's average lifetime cost per conversion is, then either switch the match type to a more conservative one (e.g. broad match to phrase match, phrase match to exact match). If it's an exact match keyword already then I'd consider pausing it.

Now, let's delete the filter we created and create a new one that will only show us the keywords that have generated more than 0 conversions (1+).

Now let's sort by conv. value / cost.

If any of the keywords have a lower conv. value / cost (ROAS) than the campaign's average conv. value / cost AND it has a higher cost than the campaign's average cost per conversion, then you should do the same; either switch the match type to a more conservative one (e.g. broad match to phrase match, phrase match to exact match). If it's an exact match keyword already then I'd consider pausing it.

Location

Making adjustments based on past performance on the location level can help to improve results as well. I always recommend targeting on the country, state, county, city or zip code level instead of radius targeting. It allows you to drill into the data to see exactly which locations are just driving up cost with little to no conversions and which are performing the best.

You essentially do the same thing as we did with the keywords; first filter out any zip codes, cities, counties, etc that have generated a conversion, and sort by cost. If any locations have a higher cost than what the campaigns' average lifetime cost per conversion is, then consider excluding it from the campaign.

Then delete the filter and create a new one to show us which locations have generated more than 0 converions (1+), then sort by conv. value / cost (ROAS).

If any of locations have a lower conv. value / cost (ROAS) than the campaign's average conv. value / cost AND it has a higher cost than the campaign's average cost per conversion, then you should consider excluding it.

Audience Segments, Demographics, Ad Schedule, Ads, Ad Assets

You basically do the same thing across all of these dimensions. This assumes you also put Audience Segments in observation mode. There are some caveats here, depending on what industry you're in, you may not be able to exclude age or gender demographics. Housing is one of those industries. Google Ads housing policy prohibits making any targeting exclusions for age, gender and zip codes.

When Campaign Performance Drops

Sometimes campaign performance drops for a given 30 day period, where you see your ROAS go down and cost per conversion go up. In this case, it may make sense reverting some of your adjustments if they have a higher ROAS or lower cost per conversion than what you're seeing in the past 30 days.

For example, lets say you had an age group 18-25 excluded from the campaign because it had a lower ROAS than the campaign's average ROAS… BUT it was generating a higher ROAS than what the campaign's ROAS is over the past 30 days then you should consider re-enabling it.