On this page you can read a more detailed description of the product model and compare its price in different stores.

In addition, shopping ads are shown on YouTube. You can see ads on the main page and in the search results.

You can order the setup of the Google Shopping advertising company in our agency. Find out the features, conditions, budget and timing of launching, optimizing or maintaining Google Shopping at the consultation:

Algorithm for displaying Shopping ads
For the user. When a user enters a query, Google searches for information on that query in advertisers' product feeds. If the information is available, the service automatically generates a relevant issue with seller ads.

A feed is a tabular file with detailed information about all products in an online store: name, cost, discounts, availability (in stock or not) and more.
For us. Shopping campaigns are created through another Google service, Merchant Center. There you need to create data feeds from which ads will be formed. When they're ready, you can integrate Merchant Center into Ads and set up your campaign as usual.

Let's go through the algorithm briefly before understanding the details:

Registration of the company in Google Merchant.
Create a product feed.
Create a Shopping campaign in Ads.
Optimize your Shopping campaign. 

Sign up for Google Merchant Center
You can start working in Merchant on merchants.google.com.

During the registration process, the system will ask you to specify the name of the store (not legal, but the one that the user should see), the website, the company address and contacts.

Next, you need to:

Confirm the rights to the site.
Add logo and store information.
Add a feed.
Add shipping and tax information.

After you sign up and upload your feed, you'll need to link your Merchant Center account to Ads. We also recommend linking it to Google My Business.

Create a feed in Google Merchant Center
In Shopping campaigns, the system automatically generates ads based on information from your feed. This is done in the Merchant Center → the Products page → the Feeds tab → the "+" button.

There are 2 ways to create a feed:

Method 1. Manually:

Google-table;
TXT and CSV files.
Method 2. Automatically:

XML file
plugins of online stores.

With the manual option of generating a feed, we recommend using Google Sheets, not TXT and CSV:

in Google Sheets, you can open access via a link and give the support service for verification;
Google Sheets does not need to be downloaded to a computer for editing, you can do everything online;
TXT or CSV makes sense if you can upload a database in these formats from the store.
The disadvantage of the manual method is that you also need to maintain the relevance of the feed manually. And if this is not done on time, the goods will be rejected. Often this takes a lot of time for a specialist, so manual creation of feeds is suitable for stores with an assortment of up to 50 on average.

To automatically create a feed, you can use an XML file that will lie at the root of the site, or connections to the API (CMS plugins). Working with plugins in the 7 most popular CMS (WordPress, OkayCMS, Shopify and not only) we explained in a separate article "How to create a feed for Google Merchant"..
The advantage of automatic generation is that it eliminates the human factor when making changes to the feed and increases the speed of updating data.
Feed Specifications
The rules of registration are called the feed specification. If you don't specify data according to the specification, Google Shopping ads won't be displayed.

Feeds include different products, each of which is accompanied by different information: ID, title, description, link, and so on. All of these values are attributes. Attributes are required, optional, and required under certain conditions.

"Required under certain conditions" attributes are, for example, a state attribute. It is mandatory for used and refurbished goods, but not mandatory for new goods. The color attribute is required for clothing and other goods that may come in different colors if they are sold in Brazil, France, Japan, Germany, the United Kingdom, and the United States. But if these products are in feeds with targeting to other countries, then the color attribute can not be used.

A couple of nuances:

GTIN – Global Trade Item Number assigned by the original manufacturer;
The MPN is specified if the product does not have a manufacturer-assigned GTIN;
If you sell handmade goods, i.e. handmade, you do not need to specify either MPN or GTIN;
the Multipack attribute is used if you sell a set of products as a single product, for example, a set of 6 candles = one product, it has 6 units;
All new feeds are moderated – it can take up to several weeks. You can find out about the status of the scan in the Products — Diagnostics tab. If there are problems with any of the goods, the service will point to it and offer ways to fix it.
Create a Google Ads Shopping campaign
If ordering the configuration, maintenance and analysis of contextual advertising in Google Ads is not your option, read how to do it all yourself and not turn gray.

To create a Shopping ad campaign, go to the Campaigns page → the "+" button → Create a campaign without specifying the objective → Shopping.

Set up your campaign
The basic settings of the campaign are almost the same as those of other types. Let's go through those parameters that do not exist (or that differ from the setting of other RK):

Country of sale – where goods are sold.
Campaign priority: If you have multiple Shopping campaigns with recurring products, the campaign that has a higher priority will win.
Product Filter: If there are product categories in your store that you don't ship to the selected country, the feature will prevent it from appearing in ads.
Local – The feature does not work in the CIS and is only available in Australia, Brazil, Canada, France, Germany, Japan, the United Kingdom, and the United States.
Location setup, bidding strategies, budget, and delivery method are the same as in other RCs.

Choosing a Shopping Campaign Structure
Campaigns with a single ad group that includes all the products of an online store at once are terrible as ineffective. The budget merges, conversions are not received, in short, nothing good. KlientBoost called it "The Mob Effect", a kind of racketeering effect, when less effective ("bad") products "knock off" the budget from more effective products. The racketeering effect is manifested primarily if shopping campaigns do not have a well-thought-out structure and settings for finding the most profitable products / queries and weeding out the least effective of them.

For example, often in a campaign with one product group, a total bid is set:

The downside is that you can't choose a bid specifically for the item.

All goods differ in cost, margin, profit. There is no reason to charge them the same rate when all other indicators are different.

Here it is worth remembering the Pareto principle about 80/20: 20% of the goods will bring 80% of the profit. Of the remaining 80%, there will be many products that pull the budget on themselves, as they are more popular in the market, they have a lot of clicks and high costs, and profits in the region of zero. So they are "racketeers".

Specialists of the contextual advertising agency when setting up and maintaining Shopping campaigns will certainly take this into account, but if you work yourself, here are two tips on how not to fall into such a trap:

bids should be set at the level of goods (product id) taking into account their effectiveness (i.e. conversion, profitability and demand);
we do not recommend automatic bidding and setting up the "cost per conversion optimizer" (eCPC) in Shopping, if you have products with different price and margins, and return in the first tip ↑
The more granular the campaign structure, the easier it is to control the situation. And the better you control the situation, the more productive the work as a whole. Especially for this, we Penguin-team have collected the most effective strategies for working with Google Shopping campaigns. First, we will discuss the main thing about them, and in the next part we will tell the practice: how to configure it, how it works, how to use it.

Importantly! A strategy is not only about structure, but also about techniques that you can use to improve the effectiveness of your campaigns, so below you'll find a couple of options for structures + a couple of strategies that you can apply regardless of how your campaigns are organized.
All products in one group with commodity bidding
With the "All products in one group with commodity bidding" strategy, we break down the "All products" product group by the ID of each product. First, assign the same bid for all products and track conversions. Based on the data on the cost of conversion, CTR and other indicators, we change the rates for each product separately in order to get maximum results.

This method is suitable for small online stores (up to 50 products) with similar products and the same price. For example, if one product is sold in several colors. If you don't have much time to optimize, use this strategy only for stores with an assortment of up to 20 products.

For example, when we were ordered to conduct contextual advertising, in one of the campaigns we set higher rates for those colors of the product that have a lower conversion cost. But if the effectiveness of the goods in Shopping is low, then he received a rate below average.

With such a strategy, you can stop spending money on products with low conversion rates and concentrate on selling the most effective items. After all, it would seem that products of different colors will have only a different CTR... But in fact, they also have different conversion rates and conversion costs. Logical: black, white and gray T-shirts are bought consistently more often than pale light green, coral and turquoise. A commodity bidding strategy allows you to use this data to improve your ad results.

SPAG — one group for one product
SPAG (single-product ad group) is the shopping equivalent of SKAG. To implement such a strategy, each product is put into a separate group.

The SPAG strategy is suitable for online stores with an assortment of up to 100 products. The difficulty lies precisely in the organization of such a campaign, so if there are more than 100 products, structuring will eat up almost all the time of the RRS specialist, and this is impractical. However, even large online stores can use SPAG, just not for all products, but only for a part of the range.

A practical example: with the STRUCTURE OF SPAG, revenue and ROAS grow, and CPA falls.

When all products are in the same ad group, it is very difficult to work on bidding, which means that it is almost impossible to improve ROAS. This is exactly the problem that SPAG solves. When you have only one product in each group, it becomes easier to increase the proportion of impressions for the most sold products, it is easier to collect negative words at the product level. The investment of every dollar in advertising is discouraged to the maximum.

You can cross-minus words. For example, with the product "moisturizer" you can take creams for the face, hands and feet in different groups and add the appropriate negative words to each group. This will help on the request "moisturizer for hands" to show exactly the hand cream.

Bidding-based structure at the query level
Shopping campaigns can't be targeted to keys, remember? And, of course, since there are no keys, then bidding on them will not work. Problem. That's what the bidding-based structure at the request level solves.

This strategy is NOT suitable for a business with a low budget. For it, you need to create 3 campaigns, which means that you can divide the budget for all of them. The budget should be enough for at least 20-30 clicks per campaign per day, i.e. if the cost of a click is on average $ 1, then the budget should be at least $ 20 per day for each campaign, and this is $ 60 per day for all campaigns. And this is the main restriction on the use of the structure, otherwise it is suitable for different businesses, regardless of the volume of the range.

Bidding at the request level allows you to segment traffic into 3 categories of requests: basic, brand, product. Just like search campaigns! For segmentation, you use priority settings, common negative keyword lists, and a total budget for all 3 campaigns. In campaigns with general queries ("shoes"), we put a lower cost per click, and in campaigns with more targeted queries ("nike air max 97 ultra black") - higher cost per click.

he highest priority gets to the overall campaigns. There are two reasons for this:

First, assigning a high priority to the overall campaign allows you to get more traffic;
second, we will have negative keyword lists with which we will distribute traffic to the relevant campaigns.
For all three campaigns, a common budget is used, because just such a step will force all 3 campaigns to participate in the same auctions. So, here the lists of negative words will manifest themselves in all their glory, directing traffic to the desired campaign, taking into account the wording of the user's search query.

Here's what the algorithm for displaying ads looks like with such a strategy:

Person enters a query

↓

Google chooses the campaign with the highest priority

↓

If there are no negative keywords in this campaign that block the query, then Google shows an ad from it

↓

If it has negative keywords that block the query, the system moves to the second campaign with the next priority level (medium/low) and shows ads from it.

With this approach, you can set the rates higher for super narrow queries and lower for general ones, because it is the narrowest queries that are more likely to buy. This means that the budget will be spent as rationally as possible: the more potentially profitable the request, the more money is invested in it.

After using this Google Shopping structure for 30 days, we got the following results:

income increased by 98%;
transactions increased by 56%;
ROAS increased by 123%.
Alpha/Beta campaigns in Shopping
Alpha/Beta campaigns in Shopping are similar in principle to a bidding-based structure at the search query level. That's just the traffic is divided not into general / brand / product, but into converting and total.

Ordering a shopping ad setup with Alpha/Beta, as with a bidding-based structure at the query level, makes sense unless you have a hard limit with a minimal budget.
 

Then we either periodically review the search queries in the Alpha campaign and add to the negative keywords those that do not correspond to the list of converting phrases. Or we install a special Google Ads script that will save our time and will perform the task automatically.

 I recommend that you read "Campaign Strategies in Ads: SKAg vs. Alpha/Beta" to better understand how these structures work.
Structure by device
Another interesting variant of segmentation is by devices: separately desktop, separately mobile, separately tablet.

The structure by device will be relevant if you have significantly different results of the RK on different devices. It will be especially valuable for you if:

You want to measure the proportion of impressions on different devices.
You want to control budgets by device.
you need an easier way to measure cross-device attribution to track a user's interaction with a brand from different devices. For example, when a user first enters from a mobile campaign → then re-enters from a desktop campaign → already makes an order from it;
you want to set individual device betting modifiers.
In general, with this strategy, we can focus on the most conversion queries. Since they will have their own budget, and we will have the ability to adjust the bid, the number of conversions will increase, and their cost will decrease.

Advice! We don't recommend combining device structure with bidding at the request level. Segmentation will be very tight, so there's likely to be very little data to be considered meaningful and rely on in marketing activities.

The exception, of course, are huge retailers with large assortments and high sales. However, you can combine device segmentation with SPAG.
Search Remarketing/RLSA for Google Shopping
RLSA are remarketing lists for search ads, originally Remarketing lists for search ads. This feature allows you to customize search campaigns for people who have already visited your site, and tailor bidding and ads to such users when they search for something on Google or on the sites of search partners. The strategy is suitable for anyone who already has traffic on the site, because to launch you need a collected audience of 1000+ people.

There are two main approaches to using RLSA with Shopping campaigns:

Bid only (now called Observations). A remarketing audience is added to the campaign and a bid adjustment is made to it, for example, +30%. This means that ads will be shown to everyone, but for those who have already been on the site, the ad will be shown in a more favorable position, since the rate will be higher by 30%.
Target and bid (now called Targeting). Ads will be shown exclusively to those who are part of the added audience.

RLSA helps maximize conversions on the search network by allowing you to bid for the audiences with the highest buying intent more effectively. The same can be done for Google Shopping campaigns with this strategy.
 
Product Bidding Strategy and FOCUS on ROI

This strategy is one of the most effective, in our experience. At the same time, there are no restrictions on its use: it is suitable for multi-brand resellers, as for monobrand online stores with any range. To implement it, ids are put into groups, as for the first strategy, for managing bets.

For example, the structure of the campaign may look like this:

group 1 — id1, id2, id3 (other ids excluded)

group 2 — id4, id5 (other ids excluded)

group 3 — id6, id7, id8 (the rest of the ids are excluded).

A single campaign can have 1 or more product category/subcategory.

For example:

* Account for children's toys.

Car Campaign:

modelki group — id1, id2, id3

"Policemen" group — id4, id5,

group "Firefighters" - id6...

group "On the radio control" - id...

Dolls Campaign:

barbie band — id...

Winx Group - id...

group "Pups" - id...

For each product, a separate rate is set based on its efficiency, price and margin. If there are not enough statistics for the product (there were no conversions), then the bid is calculated taking into account the average conversion rate. Products with very low margins and low prices should be turned off immediately.

For products with normal margins, a target CPC is calculated - the cost per click, at which sales of the product should bring the desired level of profit.

With the help of this strategy, we brought the unprofitable campaign from the client to zero (in the first month), and then to the plus.

If you used such a strategy, collected enough statistics and see that some goods still bring large losses - they should be turned off.

When working on a campaign, our contextual advertising agency focuses, and recommends you, on ROI (ROMI) - then the margin of the product and the cost of its promotion are involved in the calculation. If you focus on ROAS, as most advertisers do, the marginality of the product will not be taken into account. Which means that in terms of income we can be in the black, and when calculating profits, find out that we are still in the red.

Importantly! Like any other, this strategy requires the constant work of an RRS specialist. It is not enough to calculate the rates 1 time and apply them. Products get more statistics, the indicators change - so it is important to continue working on optimizing the campaign.
To build such a structure, you can calculate all the indicators with pens or using Excel, of course. But the time spent will often be disproportionately large. To simplify the work on such campaigns, Penguin-team has developed the Panda ppc micro management tool, which generates a report with calculations of ROI, profit for each product and Target CPC. To work, a table with the marginality of goods is loaded into the system.

Part 6: Google ads - Youtube ads (bigmoney.vip)