Google GSP ad formats have been around for some time and have been tested by a large variety of advertisers. Inventory types have varied throughout the beta, and changed along with the roll-out of Gmail’s tabbed format.
If you haven’t yet tested GSP, or you tested early in the beta and are considering it again, take a look at these tips and best practices first.
Go big or go home
Inventory is limited, with only one ad spot available in the primary tab, and two ad spots available in the promotions tab. There is no bidding down to position 3 (or lower) in this channel; either you’re in, or you’re out.
From the start of your campaigns, be prepared to bid aggressively and spend time tweaking bids to access the full extent of impressions. Keep in mind that now that this product is out of beta, there is a lot more competition; if you are not prepared to bid aggressively, have a significant budget, and accept a high cost per acquisition, GSP might not be right for you.
Bidding is CPC-based; Conversion Optimizer is available as well. Review the max CPC you’re finding success with in Display to establish a starting point. If you have CVO working well, the same principles of volume apply: more than 1 conversion/day for at least 30 days is a minimum requirement to activate and/or get results.
The more conversion volume you have in the shorter window, the better this is going to work, so consider the conversion event you’re using here wisely.
Hot tip: If you engage CVO and find that you are still lacking in impression volume, increase your CPC bid. I discovered through a few tests that these two aren’t mutually exclusive the same way they are with other Google channels.
Use proven creative
With a new ad format, advertisers will often design something new (or new-ish) to apply. Instead, use your best-performing Facebook or display (text) ad for the ‘teaser’ component of the GSP ad, and base your main ad body off best-performing email creative. As always, ensure the landing page you’re ultimately sending GSP traffic to is consistent with the creative you’ve used for both components. Additionally, start with at least 2 ads in place and maintain a constant testing regime to find what works best for GSP.
Leverage GSP’s in-form creative if you’re pursuing lead gen. This bypasses a click and moves your funnel into the GSP creative itself – it’s proven to be worth the extra effort in developing the more complex creative.
Target broadly, segment carefully
GSP performs best with big targets and minimal segmentation. I found that in optimizing campaigns by segmenting into geo tiers, I lost impression volume (even though my targets were the same). Once we tested, we were able to prove that heavily segmented campaigns gathered significantly fewer impressions than the same targets lumped together. So if you need to segment, keep it as broad as possible.
If you already successfully use GDN targets such as in-market (in GSP read: purchasers), this is a great place to start, and I experienced very similar results from those targets on both channels (thanks, Google!). Don’t be afraid to blend different targeting types together in the same campaigns (KW, purchasers, interests, placements/domains, job titles, etc.); I found that these broader groupings gave better reach and better performance, as a result of honing in on targets where several targeting types apply to one user.
GSP reporting shows performance by target, so you have the ability to use that data to optimize accordingly; start big and then weed out the bad performers.
Exclusion application is not robust
Exclusions are a sticking point with GSP. Yes, you can apply negative KW’s, remove demographic targets, interests, domains, etc., but there are a few big problems with exclusions you should be aware of before pursuing GSP. (For one, the only way to exclude converted users is to exclude those who have an email from your domain in their inbox.)
I imagine it won’t be long before we have the ability to apply a remarketing list to GSP (a great way to remove converted users at any level), but in the meantime, options are limited.
More unpredictably, containing geographic targets is a significant challenge on this platform; if you have really strict geo-targeting limitations, GSP may not be a good channel for you. Considering the “Target broadly, segment carefully” advice above, keep geographic tiers as big as possible, and remain cavalier about targeting bleed between tiers.
This channel still lacks basic time of day and day of week controls, my biggest disappointment to date with GSP. Why, Google? Whyyyyyy??? Most advertisers will experience their budgets being blown out very early in the day, which could be the worst-CVR time of day, so keep that lack of control in mind when considering GSP.
Reasons GSP should still be in Beta: (aka: my shortlist of improvements we will likely see soon)
Reporting has consistently improved, but there are still some oddities. For example, one conversion is often attributed across more than one geo target, and/or more than one targeting method. This means that decisions based on the criteria reporting need to be more holistic than you may be comfortable with, and you should have a larger delta around your CPA goal than you typically consider.
To be fair, there are some really solid data points in reporting I wish we had in other Google channels, like the URL being blocked and the % of users hitting the frequency cap.
Several basic settings we have come to expect are not currently applicable in GSP. These include:
-Time of Day and Day of Week settings
-Accelerated or Standard ad delivery
-Rotation settings for ads – you *can test ads, but achieving true 50-50 rotation has not yet been achieved in any of my tests, although this appears to be getting better and more consistent.
As with most Google releases, even those in beta for over a year, there are quirks that need to be ironed out. If you’re ready to bid aggressively and with good, well-tested creative, though, GSP is a platform that can deliver ROI-positive performance.