The Enhanced Campaign migration deadline is twelve days away. Even if you’re leaving ‘upgrading’ until the last minute, you need to be prepared now.
The biggest change will be how traffic from different devices is treated. Previously, you had control over what devices a campaign would show on – now Search campaigns* are always on desktop and tablet, and can be on mobiles with a bid adjustment. So if you had separate device campaigns, you now have to merge them, and work out what bid adjustment you should be using to best emulate what bids you want.
But don’t think you only have to worry about bid adjustments if you’re merging separate campaigns! Previously, if you targeted multiple devices in one campaign, bids on mobiles would be automatically lowered. This won’t be the case when you’re enhanced: having no adjustment means you will be using the same bid on desktop and mobile, and that means you’ll pay more for your mobile clicks.
Everyone needs to set mobile bid adjustments. So how can you work out what yours should be?
(*Display campaigns still have the option to target based on OS, device model and operator. You could theoretically have a tablet-only or mobile-only campaign if you selected the right device models, but you’d have to keep it up to date to avoid missing traffic on new models.)
Pros: Quick and easy
Cons: You will be missing out on traffic
You may not be missing out on good traffic, of course. Maybe your site just doesn’t work on mobiles; maybe your business doesn’t want mobile traffic (you are advertising PC software or Flash games, for example); maybe you’ve just found mobile traffic can’t be made profitable in the past.
Incidentally, if you haven’t advertised on mobiles before, this is probably not a good time to try: pretty much everyone will be making bid changes, so auction prices are bound to fluctuate worse than normal.
Pros: Quick and easy in the Upgrade Centre
Cons: Very probably not right for you
When you upgrade, Google will give a suggested mobile bid adjustment, but this is based on ‘similar advertisers’ rather than taking the campaign’s actual performance into account. Adobe claim that “Google’s MBA recommendations tend to be too high and mobile bids should be lowered further”. So this method isn’t recommended.
Pros: You’ve already done the hard work
Cons: Previously you’ve had keyword level bids, but bid adjustments are ad group level only.
They’ve work for you in the past, right?
If you have separate mobile campaigns then you have separate mobile bids, and you can work out what percentage they are of your desktop and tablet bids.
The problem is that mobile bids may not all be the same percentage at keyword level, and you have to set mobile adjustments at ad group level.
If there’s a really significant difference you could split up the ad group. However this makes management more difficult: while you want small ad groups for better ad relevancy and more precise bid adjustments, in some cases ad groups don’t have enough traffic to be worth the extra managerial effort.
So in most cases you probably want to use weighted averages – combine the keyword level adjustments together, but giving more weight to the adjustments from the keywords that do most of the spending. Tomorrow I’ll go through how to work this out in Excel!
Pros: Fairly straightforward
Cons: May not reflect mobile traffic’s actual value. CPCs are not the same as bids.
But what if you haven’t split out your campaigns?
In legacy campaigns, Google automatically uses a lower bid on mobile searches, so your historic CPC on mobile should be lower than desktop and tablet. So if you just want the status quo to continue, you can calculate
(Mobile CPC / desktop and tablet CPC) - 1
Again, there may be variance at keyword level. Again, you’ll want weighted averages, which I’ll talk more about tomorrow.
This method is fairly easy – it just needs some Excel work – and you can see your current mobile performance to give you an idea of how it will perform. But it doesn’t take into account conversions: you’ll carry on as before rather than getting any improvement.
Also, it raises the problem that CPCs are not bids: lowering a bid by 20% could reduce the CPC by more, or drop you off the first page. It depends on your competition.
Pros: The way you calculate your desktop bid is probably the best way to calculate any bid
Cons: It’s effort. Average position is more important on mobile. Again, you can only adjust at ad group level.
(Value of a mobile click / value of a desktop and tablet click) - 1
‘Value of a click’ could mean revenue per click or conversion rate, depending on whether you have revenue tracking in place.
This assumes your bids are set up according to value per click. You might instead be bidding for position, for branding purposes. Or you might be bidding less than your value per click so that you can get the most conversions from your limited budget.
There is also the complication that mobile SERPs have fewer ad slots: higher position is more important and you might want to bid a bit higher to actually be visible.
But basically the idea is: work out what the mobile bid should be, as if you were just targeting mobiles on their own, the same way you would for your desktop bid. Then find this bid as a percentage of the desktop bid. You could do this at ad group level, or work it out at keyword level and change that into an ad group level adjustment the same way as in method 3.
I realise putting “it’s effort” as a con sounds lazy, but if your process for finding the best bid is manual and you’ve got too many keywords then there’s only so much you can do before the 22nd. You might only have time to work out the right bids for the most important campaigns, and use adjustments based on historic CPC for lower volume campaigns until you have time to work things out properly.
Pros: You can fine-tune bids more finely
Cons: Can’t predict the effects precisely.
But wait! What about location bid adjustments? They will layer with the mobile bids – if you’re using both you should probably check you’re not throttling or exploding bids more than you want to in any particular location/device combination.
Example: the people of Townsville have low conversion rates and high cost per conversion, so you drop bids there by 20%. And mobiles aren’t doing well either, so you also drop them by 20%. What if it turns out that the reason Townsville isn’t doing well is that it’s full of mobile users? You’ll effectively be dropping bids by 36% there, because the adjustments multiply – you’ll be bidding less than what you actually want.
It’s unlikely that traffic somewhere is 100% mobile, but the proportion will vary by location. While you can’t get a precise prediction of your adjustments’ effects, you can make educated guesses – you can use a User Location report, segmented by Device, to see what the proportions are and what sort of overall adjustment will be made. This is another job for Excel – I’ll go through this on Friday.
However you calculate your mobile bid adjustment, remember to monitor it! With all the changes of enhanced campaigns, the prices are bound to fluctuate when other advertisers try out their own bid adjustments.