Next year UK Merchants are going to have Valentines Day with a difference. As many will know Google is making one of the biggest changes to its search results landscapes in a long time, with user experience being the key driver.

On February 13th 2013 the migration to paid for traffic will begin. It won't be an instant switch as with most things Google do. They have also just published an incentive for merchants who create PLA campaigns for all products by April 13, which could be upto 10% of monthly PLA spend. The UK will be next but Australia, Japan, Brazil, Spain, Italy, France, Netherlands, Germany and Switzerland were set to follow suit early in 2013, with full role out complete by end of Q2 2013. These markets are already set up for the new PLAs so you can roll out activity in all key markets now.

The main drivers cited by Google are:

• Poor Quality data from merchants (due to FOC traffic)
• Difficulty in policing incorrect information and misleading merchants and/or poor service (without it being paid its difficult to give adequate resource)
• Challenges for merchants in terms of differentiation vs competitors and control of traffic levels

In addition to the standard SEO listings and the traditional text based AdWords advertising Google has been introducing new ad formats and included more sources, such as Google shopping, in the results. This often leads to a very crowded results page, with elements of duplication in some circumstances (like the example above).

The key elements highlighted are the areas driven by the Merchant Centre feed and will be affected by the changes or impending changes.

1. Product Listing Ads (PLA) - these are driven by the data feed and must be linked to a CPC bid by creating a campaign in the AdWords interface. Currently these cannot be controlled, Google will derive the most optimal match from the merchant centre feed
2. AdWords Extensions - These link products from the Google merchant feed allowing the user to further drill down their search, the biggest difference is that these are defined at keyword level.
3. Universal Results - sometimes referred to as the "one box", is the area currently where shopping results will display as well as other relevant search results. This is the single biggest driver of traffic to merchants and to the Google shopping property itself and has led to Google’s meteoritic rise to become the most visited shopping comparison destination in 8 global markets.

New Google Landscape with the single PLA section

The biggest change on the Google landscape, is the merging of 2 distinct sections of the search results pages, namely Product Listing Ads (PLAs - not to be confused with extensions) and the universal product results (the “one box”) into a new single section - effectively PLAs on steroids, referred to by Google as the commercial sponsored unit or “the unit”. This change is already live in the US and this is being rolled out to regional sites throughout this year. The bidding for this is controlled through the existing AdWords interface and data is powered by the Merchant Centre data feed.

The days of thousands of free clicks from Google shopping are gone, as they will now be charging for it. For some this is an opportunity if done intelligently and efficiently, for others it makes selling more costly and difficult. All the old elements of Google Shopping ranking algorithms will remain, so just like any of the other Comparison Shopping engines (CSEs) where you rank will be determined by a combination of:

• Relevancy
• CPC bid
• Popularity and CTRs
• Price
• Customer feedback
• Stock position
• Delivery cost

In addition Google have access to all the product attributes via the feed you submit and data from elsewhere which are likely to make their way into the mix in the future. We have already seen examples of merchants being delisted for poor customer service this year.

What does this mean for me as a merchant?

There are many burning questions such as....

• What is the best strategy to be successful on Google shopping and in the new PLA environment?
• Who is best placed to drive this strategy for me?
• What technologies should I be employing to help me make the most of the opportunity?
• How much is it going to cost me?
• What should I be bidding? How should I be bidding?
• Should I list all products, if not which ones?
• How do I make sure my data feed is optimised and meets Google’s specification?
• How can I optimise the content of my feed further to give me a competitive advantage?

We are going to try and address some these questions and concerns in this paper.

When should I be using PLAs

PLAs even under the current system are proving to give excellent returns for merchants, much higher CTRs and conversions, you can read more about this in a recent article on E-Consultancy [http://econsultancy.com/uk/blog/11018-google-product-listing-adsdeliver-higher-ctr-than-search-ads]. If you've not yet run any PLA activity then set some up soon, to give yourself a little bit of learning before the new PLAs come in.

Getting the basics right first - make sure your feed is awesome not just compliant.

The most important thing you need to do is make sure your Google Merchant data feed is optimised fully to Google’s spec and ensure it doesn't just meet the basic requirements. You want to go the extra mile add in everything you can, optimise your titles and descriptions. Remember this is like SEO in a microenvironment so like any other search channel, if you sell trainers and none of your titles contain the word trainers they won't show up! You can read more about how to optimise your feed for Google on our blog article, Google Product Search 101, written by our Integration Director Matt Sullivan. We will be following this up with a more detailed dive into optimising for Google Merchant centre.

The listing in Google Shopping results and the new PLAs will be powered by the data feed just like they are now with the Merchant name, price, product title and pretty much all of the product matching data coming from the feed. Google will just look to utilise more product attributes in the new PLA units, to enhance the users filters and thus help them find products better.

Make sure you use the promotional text wisely, and include it in your feed, applying logic hierarchies if you are using systems which apply business rules i.e. a product may be in a brand that is offering “Free Shipping” but the specific product qualifies for “30% discount till Sunday”.

Remember to setup the relationship between your AdWords accounts and your merchant centre feeds. Linking your feed to your PLA campaigns is vital as this is what drives all of the content of the PLAs. Currently a secondary PLA feed can be utilised in the Merchant Centre, which is less detailed than the main Google feed. This approach is not advisable as it limits the potential of the campaigns and also these secondary feeds will soon be disabled and everything will have to run through the Google Merchant feed. Indications are that this will be around March 2013 so it is critical to have the optimal approach implemented sooner with all the impending changes.

Manually creating PLA and shopping campaigns

1. Submit you feed in the Google merchant centre ensuring it’s exactly to spec
2. Add groups to your feed if you are going to try to manually create campaigns, it will help the process, companies with tools like ours can do this dynamically for you if needed.
3. Link your AdWords accounts in the Google merchant centre – go to the settings section on your Merchant Account centre and link it to your AdWords Account ID
4. Add a default bid and daily budget in Merchant Centre
5. Create your PLA campaign in AdWords using the create new campaign wizard
6. And add the targets in the ad group view
7. Campaigns can be created using a number of attributes from the feed such as brand, category and product ID, you can also use labels or groups if these have been built into the feed.
8. Once you have created your Campaign and AdGroup, you will need to create an Ad. Once you have opened your AdGroup, click the New Ad option and select Product Listing Ad. You can also define your promotional messaging here.
9. The final step is to enable Product Extensions for the Campaign by going to the Ad extensions tab, selecting View: Product Extensions and then creating a New Extension. Please ensure that you select the correct feed from the Account drop down as this could significantly impact your activity if not correct.

Dynamic creation and updating of all PLA campaigns from your feed

Intelligent Reach’s system uses data feeds to create any level of campaign and update this on an ongoing basis; we use the same technology to generate all of the PLAs at product level, which are then pushed to Google AdWords via our API. We will normally build the structure of the campaign to replicate your site categorisation just like we do for product long tail campaigns.

So how do I manage my PLA campaign?

Here's the biggest change from tradition AdWords Campaigns - there are no keywords. The new PLAs and shopping results are unlike anything else managed in AdWords or many of the popular bid management platforms. In the new system you bid on your products to show when searches relevant to those products are performed. You can decide bids based on brands, groupings or even specific products.

Which Products should I list and if so how much should I bid to start with?

If you have been tracking your sales activity in Google Shopping then you should be in a good place to know what products can afford to be listed at what bid. For many of our merchants we have over 6 years’ worth of sales data down to a SKU level so it’s easy to reverse engineer the bids that can be afforded. Its important to include as much other information that you can, i.e. bids for your products on other shopping sites, long tail PPC campaign bids, etc. This provides a very good starting point for any product level bid strategies. This can also be more generalised to group bidding if so inclined. The pros and cons of this will be discussed later.

How much extra budget do I need to find for this?

The general rule is that the more budget you spend the more return you are going to get. At first you should monitor your budget closely, and have one that you feel comfortable with. Over time as you get more data and find out what products sell well and what don’t you should be able to remove those poorly performing and increase the bids on those that perform well. A highly optimised and automated PLA system shouldn’t need a budget cap as it will be running profitable, so the more spent the more real profit gained.

However traditional large organisations have to set a budget for the expected costs. Intelligent Reach can help give you a good idea of the exact budget based on current Google activity, which when combined with current PLA spend should give an accurate starting figure. How much you spend depends on how aggressive you are at utilising the channel, how strict your margin targets are and how much consumer demand increases for the new shopping experience.

What Bid Strategies should I employ?

Generic Bidding

The traditional way to bid in Google AdWords is to bid at a generic level. This makes sense when bidding on keywords however it needs a lot more thought when you are bidding on products. In our experience running Generic and Long Tail campaigns is that they need to be treated very differently and with very different strategies. Generic campaigns often run by comparing the costs with a conversion metric such as Revenue, and with volume this is an easy metric to use.

However with Long Tail the keywords are associated to a particular product, normally very specific, with good natural quality score. In these situations the traffic is low, but conversions are high, it is important to look at other metrics to make optimisation and bidding decisions, such as price competitiveness, seasonality and performance vs specific product margins. Long tail is closer in its bid strategies to those of traditional CSEs.

The key question now is, should you apply strategies well honed in the CSE space or those developed and working efficiently in the AdWords environment. The most important fact here is that there are no keywords to manage, the optimisation is around products, displayed based on Google’s interpretation of your feed content (the relevance and data Quality), as well as quality score, CPC bid, popularity, CTR and all the other measures Google puts in place.

No one can determine the exact mix or priority of any of the ranking algorithms out there, whether on a CSE like Shopzilla or in Google’s new model, however the most important fact will always be relevancy, else they run the risk of alienating the user.

Group Bidding

Google is facilitating a range of bidding options around building groups based on the below variables (from the merchant centre feed)..If you are using other systems you may be able to build up more complex grouping structures.

• Product Type (or subcategory)
• Brand
• Condition

If you have a provider or your IT team have the right tools you can use additional fields in the feed. The later of these is better to use, as it can accept multiple values, so products can belong to more than one group (with the highest bid taking precedence)

• Adwords Label
• Adwords Groups

The most important thing if you are going to employ this strategy is to set an overarching product catchall bid at a low value for the Ad Group. Negative Keywords - same as AdWords this ensures your PLA does not show for any non-related search term, so if you are only selling leather jackets you can list of the other jacket materials such as denim jackets, fur jackets, sports jackets etc. Using automated advanced search query reports can automate this process. It’s very important for generic campaigns, less so for granular ones.

In some situations it makes sense to group products together to bid the same, for instance if you have a sale you may want to bid up all products in that sale by a similar amount, but does it make sense when each of these products has different margins, different price points, conversions traffic levels etc. You may want to bid them all up by 10% say but not blanket bid them to 85p, surely?

Product level Bidding - no minimums it’s utopia!

Intelligent Reach believe that the company which manage and (hopefully!) optimise your data feed must be at the heart of any strategies, we believe that although it makes the most sense for Google to house this new opportunity in AdWords, it doesn't have as much in common with generic keyword bid management as it does with CSE management.

CSE's are given a hard time when they are actually trying to do something pretty complicated. It’s a massive juggling act, arguably with one hand tied behind their back. They rely on merchants supplying good quality data, which they can then ingest and display to their users to help them find the products they are looking for, adding value to that user simply visiting some key merchants direct sites, adding in price or feature comparison, stock levels and other buyer’s experience. They need a certain number of merchants selling each product but their CPC rate card and traffic mix (and thus conversion) can hamper this. Its chicken and egg, they need offers but the merchant at the minimum category bids need buyers else they stop listing their products, perhaps then the CSE has finally got its SEM buying platform in action and is buying good quality traffic i.e. Buyers but there are now no merchants selling the product.

Google has got round this by offering the clicks for free, meaning every merchant lists every offer, they invest in the data optimisation as a point of differentiation and get decent ROI on this investment as there are no media costs. The Google Merchant feed is now also so integral to so much else that they do, to the point where other CSEs and other platforms are working to accept the Google format as standard.

The solution that so many CSEs cannot or will not implement is product level bidding with no minimum bids, or a hybrid CPC/CPA model. If they need inventory coverage, the simple fact is that operating with a rate card with minimum bids at a subcategory level means that product level optimisation systems like ours have to remove products which don't have a low enough cost of sale to achieve the merchants target, and whereas if there were no category minimums we could bid it down to its "sweet spot".

CSEs, which offer the most flexibility, are sites such as Shopzilla and Nextag, who whilst operating subcategory level min CPCs, do allow bidding at a product level and as such they retain a lot higher % of merchant SKUs in general. Perhaps the most flexible CSE is Pricegrabber, which has only guide subcategory minimums, so we can set our bid management systems to find that sweet spot right down to 2 or 3p.

AB Testing feed content for optimal results

So depending on the consumer search term, Google is going to show the most relevant product, which they think the consumers are showing interest in. So for very specific search terms very specific products will show, optimising what appears on more generic searches is much more difficult and challenging. It requires a long-term strategy of feed content optimisation, in terms of title and description content mainly, and then a system which can track changes and their impact on visibility (clicks and impressions) and performance (CTRs, conversion %, AOVs, sales - in the different attribution views) for each term. It’s like SEO in a microenvironment, and just like SEO focusing on one keyword can be to the detriment of another.

It doesn't always follow that the products, which look the most relevant for consumers, are actually the best ones to promote, its common place for CSEs to have what are termed "runners" - products that make their way to the top of search rankings but never convert, you may see this with products that have broken URLs which don't resolve or incorrect feed information such as price. These often need corrective action or removing all together.

Quality Score, Trusted Store status and other points of competitive differentiation

The trusted stores programme is currently only available in the US (www.google.com/trustedstores) however this will soon be launched in the UK and discussions are ongoing between Google and key UK retailers.

The importance of intelligent reporting and automated decisions

Whilst many platforms, agencies and merchants have struggled to track the performance of Google shopping over the years (due the fact that redirects are not allowed), we at Intelligent Reach have been building up over 6 years’ worth of performance data for our merchants across a broad range of product areas. For our existing merchants this means we can reverse engineer all your bids at a product level the second Google shopping starts to move to paid for and we are also working on our own bidding guidelines at a category level Google themselves are working on tools to help forecast this but in the meantime we hope this resource will be valuable

The Intelligent Reach platform gives a wealth of reporting and actionable insight to our account performance teams and merchants, including intelligence around contributing touch points, price changes and price competitiveness. Our performance algorithms allow our users to automatically manage 100,000s of PLAs, long tail keywords and generic terms across a wide range of sites not just Google. We can help you understand how users are interacting between these 3 sections of your campaigns, advising which are more effective in which combination for which kinds of products.

New Enhanced Campaigns – weight your campaigns towards different devices

Google are now introducing another significant change in their service, which will have a big impact on how you optimise your adwords, PLA and Google shopping campaigns. The new Enhanced campaign feature allows you to manage cross device activity. It effectively allows you to add weightings for users on different devices at a campaign level, adding uplift or down weighting applying % change to your bids. In addition to devices it also allows similar weightings to be added for Location and time of day. We will be covering these changes in our blog as soon as we start some test activity.