Episode 19: Google Shopping support and automated bidding algorithm changes

Play this episode:

Or get it on:

• Apple iTunes
• Anchor
• Breaker
• Google Podcasts
• Overcast
• Pocket Casts
• RadioPublic
• Spotify
• Sticher

What's in this episode?

In this episode Mark Williams-Cook and Rob Lewis will be dicussing some recent news which has affected Google's paid advertising platform. First is the recent new ruling from the EU Commission which is prohibiting Google's support of their Shopping platform. Secondly is a change Google have implemented to their automated bidding algorithm, which could have positive effects on SMEs.

Show note links:

European Commission ruling press release: http://europa.eu/rapid/press-release_IP-17-1784_en.htm

Transcription

MC It's Monday and it's time for another Search with Candour podcast, my name is Mark Williams-Cook and today I'm joined by the Grand Master of ads Mr. Rob Lewis.

RL: Hello!

MC: Today we've got a bumper PPC episode for you!

This is episode 19 of Search with Candour podcast recorded on Friday the 19th of July 2019 and we have a PPC bonanza episode for you where we're going to be talking about Google's €2.42billion fine for Shopping ads and what that means for you. We're going to be doing a deep dive into the major changes they've made to their automated bidding which is very likely to affect you, it hasn't been spoken about much at the moment and Google’s only been emailing people about it.

On the 27th of June 2017, so we’re talking over a couple of years ago now, the European Commission fined Google a whopping €2.42billion for breaching EU antitrust rules and they were saying “Google has abused its market dominance as a search engine by giving an illegal advantage to another Google product in its comparison shopping service.”

Commissioner Margrethe Vestager, who was in charge of competition policy at the time said:

We know Google Ads is obviously massively important to Google as a company with around 90% of their revenue coming from these ads and we've just been told in the last few weeks that Google is confirming as part of this ruling they're no longer going to be giving customer support for Google Shopping ads and moving forward they'll only be offering what they call “best practice guidance”.

Google issued this statement; so originally actually they were emailing Google partners and when we're having our monthly calls with Google they were saying that they are stopping this support, but who will also issue this statement saying:

"Following the European Commission’s June 27, 2017 decision, Google has made Shopping ads on Google general search results pages in the EEA and Switzerland equally available to all qualifying Comparison Shopping Services ('CSSs'). This is directly impacting the support model for Shopping ads. Our team and Shopping teams are no longer able to provide support for Shopping campaigns. This includes any optimization, implementation and product feed troubleshooting relating to Shopping campaigns. From now on, we are only able to share best practices and new product information for Shopping."

So, Rob I think that this seems like it would have, I guess, probably the most impact on SMEs who maybe aren't using agencies at the moment where perhaps they don't have Google Shopping or Google Ads experts, or maybe they don't have technical teams who understand the ins and outs of the feed implementation. What do you think the impact is going to be here?

RL: I think there will be quite a big impact in terms of, well, having a lack of technical support is going to be a problem. There's been times where I've had to speak with Google's technical support team in order to fix an issue for which there was no documentation because it had just never come up before. There's always something really unique and niche that causes a feed to stop working that requires Google to investigate the issue and come back to you. That's happened to me so many times and quite often it's an error on Google's part where their automated systems have incorrectly marked a feed as non-compliant, and it does happen and in all of those cases I've had to speak to someone at Google to get the issue investigated and for them to fix it their end or they tell me what I need to do to fix it my end. There are always technical things that can be fixed your end as well but I guess we aren't having that support, it will be a problem.

For us as an agency it's not the end of the world because we've got many years’ experience of these things happening and we can draw upon our previous experiences of when things go wrong and fixing them but as you say for SMEs it may be a problem. There is always an issue with a product feed, at some point at least once a year every feed will get a nasty warning from Google to say that it's non- compliant for one reason or another, it does happen and a lot of the time it's due to Google misinterpreting the information that you've got there and its automated systems flagging it as a problem and it's not. But in most cases that can be fixed by speaking with Google.

MC So is that always necessary? I mean again as an agency, we have a luxury which is that we have a direct contact with Google for this and again that's not going to be something all SMEs have. They're not going to have this person that deals with their account and it's probably only going to increase the volume at which people are going to need to contact Google.

RL: I would say so. One of the things that's worth mentioning about Google Shopping feed disapprovals is that when your feed has been flagged as non-compliant they normally give you a 30 day grace period to fix it, and it used to be that once you've fixed it you contact them it will be manually reviewed and then they'll come back to you with anything that needs changing. They actually changed it last year so that from now on when you fix your Google Shopping feed you don't tell Google to take another look at it, they just say after the grace period of 30 days someone will look at it and someone will confirm if it's acceptable or not. If it's not acceptable it will get shut down.

MC: Okay, so that means that potentially you could fix an issue I guess in that 30 days, or rather believe you fixed an issue.

RL: Absolutely.

MC: And you have to wait until the end of 30 days, and then there's no extra grace period on if you've done it wrong.

RL: Yeah, they'll go “no that's wrong, switch it off” and then you have to go through your other channels. I don't know what's going to happen with those channels if they're removing all technical support. I'd like to think that there will at least be some skeleton crew, so to speak, that will manually review the policies but by the wording in the statement they've given it suggests that there's not going to be any guidance to do so, specifically for things that fall outside the best practice or isn't for whatever reason covered in the standard documentation.

MC: This looks like from the statement it's effective immediately?

RL: It says, that's what it looks like yeah, I mean they haven't made an official statement as of yet I think. They’ve only just been contacting agencies to tell them about it so I have no more information on that at the moment.

MC: Okay, so takeaways?

RL: Make sure your shopping feed is compliant because the last thing you want is one of those nasty Google emails warning you that it isn't compliant, and also just make sure you've got access to someone who has experience working with feeds. You might be using a plugin for your feed - you may have approached someone to create a feed or you may have created the feed yourself using Google sheets or some other type of spreadsheet.

All of that's fine, just make sure that whoever's managing it is keeping an eye on it and has the ability to edit it quickly if it needs an update. The last thing you want is your shopping feed to be shut down during your busiest time of year.

MC: Automated bidding! So, I love talking about automated bidding anyway, one of my interests as you know is the whole AI side of things and Rob and I have got differing opinions sometimes when it comes to automation strategies, but this is a really interesting bit of news we've had because this is something again I believe Google haven't put out. I haven't found anything that they've put out about this publicly yet, which is that they have made a fairly substantial update to their automated bidding algorithm. Do you want to just take us through that Rob? The changes, it'd be helpful maybe just to talk about for those that maybe aren't using them or are a bit rusty about automated bidding strategies - what's available and what they're used for.

RL: Well typically when you're managing a pay-per-click account you'll have two immediate options: one is to automate your bidding strategies and the other is to manually control them and set your own bids, and there is also, just to be pedantic, somewhere in between where you can manually manage your bids but add some element of automation in order to help - that's the enhanced cost per click stuff.

There are different types of automatic bidding options that you can choose from, some will be more relevant to advertisers than others. To give you a brief overview of the main ones:

The first one, and it's normally the one that Google pushes for you to use when you set up a new campaign, is the “maximise conversions” bidding strategy. Now the way that works is that Google will try to spend your entire daily budget and do so to get you as many conversions as possible.

MC: That sounds like a good thing, on the surface!

RL: Maybe, so just to put that into perspective. Let's just say you have a £100/$100 per day budget Google will do everything it can to spend that budget in order to get you as many conversions as possible. Now it might be that in any given day you would only get one conversion, Google doesn't really care - it wants you to get that one potential conversion but at the expense of spending £100. That might be really good if you're not too concerned about the costs per acquiring that conversion.

MC: So, if you're just trying to grab market share for instance and you're not that worried?

RL: Yeah if you've got a daily budget of £100 and you just want to get at least one lead or as many leads as you can every day out of it and you're not too focused on the cost per conversion, maximise conversions is fine. That's just something to consider.

The next one that relates to automated bidding and conversions is “target CPA” which stands for “target cost-per-acquisition” which is slightly different to maximise conversions because you tell Google what you want your average cost-per-conversion to be and it will do everything it can to meet that. I suppose this is better for lead generation campaigns where you know that each lead is worth a certain value and you know that lead is only going to be profitable if it's no more than a certain amount of money for you to acquire that. For example you might want to set your target cost per acquisition to £20 per lead, knowing that most leads are worth several hundred pounds and that you close X number of leads that you receive. So, target CPA is a good way of I guess having some kind of expectation as to if the lead is going to be profitable for you.

The next one is “target ROAS” or “target return on ad spend” which is slightly similar to target CPA except it looks at the actual revenue figure. So, this is for e-commerce advertisers - that's what they're focusing on, sales, specifically sales that they want to make sure they generate a profit for. In such circumstances you don't really look at the cost per conversion (it is as an important metric) but actually it's whether or not that conversion is generating a profit for you.

MC: So would this be if you're selling a range of products, some are £5 and others are £500?

RL: Yeah absolutely. You might be happy to, let's say your average cost per conversions is £10 but you have a specific product that sells for £1,000. You might be happy to sell that at a cost per conversion of £50 because you know you're going to get a good return on that spend from it. So, target ROAS you set a percentage, you might set a target ROAS of 500% and Google will do everything it can to get you that figure.

Another one that they're pushing a lot lately is “target impression share” which is slightly different, it doesn't really factor in conversions but instead what it does is it allows you to choose an ad position that you want to achieve, such as the absolute top position or somewhere in the top three, before ad positions and choose a percentage of time that you want to appear in that spot. If you wanted to appear 100% of the time for specific keywords in the absolute top position, you could set it to 100% and Google would do everything it can to ensure that you achieve that position.

Of course another automated bidding strategy is “maximise clicks” which doesn't focus on conversions or anything it just does everything it can to use up your budget and generate as many clicks from that budget as possible.

So. there's lots of different automatic bidding strategies.

MC: Where does manual bidding fit in, in 2019? Because a lot of these sound great. Is there still a place, do you think currently and in the future, for manual bidding?

RL: In my experience yes, automated bidding still can't always reach the results that a manual bid can achieve and there's various reasons for that and that's such a long conversation to have - I can talk and talk and talk for a long time as to why. It might be worth talking about the actual update itself to the algorithm, then I can come back to that.

MC: Okay let's do that.

RL: Okay, so historically Google's automatic bidding strategies - the ones that focus on driving conversions, historically they relied on previous conversion data in order to optimise your bids. It looks at conversions for the last 30 days and it would see where those conversions came from and it would build on it and it would update the bids in order to try and get more of those conversions.

MC: So it's just using your old existing data to basically train a model of ‘this is what's working’?

RL: That's right. If certain keywords showed that they consistently drove conversions and someone typed in a search query that matched that keyword, Google would say ‘historically that search query has performed well so I'll increase the bids on there and reduce the bids on the searches that haven't’ so thus it removes the manual need for you to manage those bids.

Now in the past you had to have a certain amount of conversions in your account in order for those bidding strategies to be available to you. Now, Google have just confirmed - that the algorithm has had a core update at the account level and they have said “we are using new algorithms at account level, based on audience signals from all advertisers within the same vertical” and they also confirmed that you no longer need conversion history in order to use bidding strategies.

Just to step back for a moment, you can set your bidding strategies and maximise conversions or to target CPA or target ROAS without having any historic conversion data because Google is going to look a range of signals and it will determine the intent of the user in order to decide which searches it should increase bids for, and which it should reduce bids for.

MC: So, actually they're kind of taking all your, what will likely be competitors’ data, aggregating it and sticking it in a black box and using that intelligence to help you optimise a campaign?

RL: Absolutely, I think it's a great step forward that they're looking at signals from other advertisers.

MC: I guess it lowers the bar in terms of, you know when we speak to new clients or people who are just starting paid search. One of the biggest hurdles I see for new businesses who want to do paid search is that their competitors have maybe been running ads say for two years. They know which keywords work, they know which ads work, they know which landing pages work, they've optimised their cost-per-sale, cost-per-click so it's a huge investment sometimes just to buy yourself the intelligence so you can compete on that level because from a fresh entry you're not really competing with those PPC veteran companies. You're basically buying some intelligence so you can get to the stage where you compete. This kind of shortcuts that in a way of Google saying, ‘we will use some of that data to work out what's going to work for you’.

You're kind of taking investment, Google's giving you some of that investment from other people.

RL You could look at it even deeper actually and say that for all this time Google has been getting all the other advertisers to figure out what works best and that information has now become available for it to just completely automate everything, a fully automated bidding strategy based on what has worked for every single advertiser before that earned conversions - just press go. They’ll automate everything because this is what's worked for worldwide advertisers that use the Google Ads platform.

MC: And that's information obviously they're never going to share with companies or be able to share because of privacy reasons which is good because it's convenient, because you can resell it.

RL: Yeah absolutely, and the entire thing I think it's a great step forward. First of all, it's also an interesting one because historically Google has said that its conversion bidding algorithms for automated bidding already used lots of signals and didn't just rely on conversion data. In my personal opinion, which doesn't reflect any agencies for whom I worked with, just to make my personal opinion is that the update is now changing the bidding algorithm to how it always should have been and how I would always have wanted it to have been, which is a user-focused approach one that looks at the intent behind the search and doesn't just look at historic conversion data.

I'm really looking forward to seeing what positive impact it hopefully has on automated bidding strategies that we're using at the moment, and you said Mark that you and I sometimes have different thoughts and feelings about automated bidding. I've always been very much of the opinion that if it's Google's automated bidding solution it's probably the best one out there because it has access to so much data and when it comes to non-Google affiliated tools, I'm sure a lot of advertisers got contacted by non-Google affiliated tools that offer awesome automated bidding strategies, and I've used a few of these tools before just out of interest and they're all targeted at the layperson. They're all about simplifying things and generally speaking they don't have access to the vast amounts of data that Google has, so I've always been very sceptical about using them for my clients’ accounts, I’d much rather rely on Google's vast reams of data about its users.

MC: I guess there's that balancing act of Google's definitely got the most data which they won't be able to share with people who have their party tools but there's the motivation I guess which is maybe a third party tool probably has a motivation closer to the client in that they want to make you money, so you keep using a tool whereas a lot of Google's moves we've seen - well they've had almost a some type of Halo/positive effect I'd say on advertisers. The primary driver for their changes and actions is their revenue, their profit and their growth.

RL: A very good point

MC: Which is why you know it's no secret within our industry those that have Google Account Managers, it's a very common suggestion that to improve your Google Ads account you should increase your budget. That's a thread that comes up quite regularly when we speak to people and Google.

RL: Yeah absolutely, and just to clarify I've always been a proponent for Google’s automated bidding strategies but unfortunately it has to be said I've always got better results when I've controlled bids myself, particularly on the most important crucial traffic; every area of the pay-per-click campaign I’ve seen Google's automated bidding has never managed to match just human intellect in my opinion but not yet. Perhaps this change is the change that we've all been looking for - the one that's going to make it even better now. I think it's moving in the right direction because one of the things I've been seeing more and more over the last 10 years is that less keywords consistently convert, that there's fewer keywords that comprise what I would call the “core keywords” which comprised the vast majority of conversions for a client.

MC: It's now the longtail stuff, it's what we're seeing in SEO as well which is that several years ago we would target, we'd pick some set of key phrases that represented a large proportion of our traffic, whereas now with a lot more focus on basically writing better content now and that's naturally becoming more diverse and people are seeing that the traffic they're getting is more diverse in terms of the searches. Users are now being more specific with their searches. I think we get better search results now when you're specific.

RL: Everyone searches for what they need which is unique to them and Google of course personalises the results for what they've previously searched for and makes assumptions as to what that person is going to be interested in.

So the entire search results for everyone is unique to them and just to use a really bad analogy, someone could Google the search query “awesome sunglasses”, see a pay-per-click advert of awesome sunglasses, click on the advert and purchase those sunglasses. Whereas another type of person might type in exactly the same search, click on the advert but not convert so it's the intent behind the person who wrote that search query that matters and that's what Google is trying to factor in with this latest algorithm and it makes me ask the question: ‘How important is the search query, moving forward?’ which again is probably a discussion for another day.

But if the targeting that Google can come up with behind the intent of the user is so accurate then we may get to the point where we can target people by their exact shopping wants and needs regardless of what they're typing in, which is what Google is moving towards and in theory they've provided us with tools to do that now but it's still not 100% accurate. So, someone could type in “mobile phone tariffs” but we could show an ad for sunglasses and they may buy those sunglasses because the intent targeting was so accurate.

MC: Okay, so this is actually moving away from any keyword-led targeting to the moment where we're at the stage then where we're looking at keywords in context of intent. So to take your example of mobile phone tariffs there might be intent as to whether someone's actually looking for a new mobile phone, or if they're all just doing a completely unrelated bit of research because this search is unique to the user.

RL: Google's new conversion algorithm should in theory factor in more than the intent behind a signal, and of course it will look into historic performance amongst other advertisers and it will then decide whether to bid high or bid low, so it's factoring in a lot more about the user rather than just the keyword, rather than just the superficial searches that we’re bidding on.

MC: So, the big question here that we've skirted around and I want to ask is, how long until we get to the stage then where we don't need PPC account managers?

RL: Well, put on the spot now I think personally it depends how much control that Google allows the advertiser to have. If they provide all of this automation then but they also provide the means to override the automation and do what the advertiser thinks best which I would like to think they do. So, then I think there will always be the need for an advertiser, always needs to be a consultant that analyses the data and decides where best to direct money, which numbers to influence in order to make the best long-term goals for each client.

I mean Google doesn't understand the goal, it doesn't understand the human aspect it only understands the numbers and if you focus too much on the wrong numbers then things can sometimes go a bit haywire. I'd like to think that there'll always be a need. But, that's I guess that's down to Google isn't it and how much control they want the advertiser to have.

MC: That is a really great answer and I expect lots of PPC managers to repeat that verbatim to their boss.

That's everything we've got time for in this episode, we will be back on July 29th next. All of the show notes, links, transcription as usual are available online at search.withcandour.co.uk, I'm Mark Williams-Cook I hope you have a wonderful week! I look forward to hearing your questions, and we will speak to you next week!