Before getting into what our search accounts look like, we should explain a little bit about why we structure our accounts the way we do. For one, having an organized and clear structure allows for easier navigation – you save a lot of time when reviewing performance. It also leads to more precise reporting, which makes you better equipped to optimize where needed. A well-structured account is the foundation for succeeding with your paid ads.

The “Perfect” Structure‍

To be clear, there is no one-size-fits-all structure for a search account. The “perfect” structure is the one that performs best for your business, is in line with company goals, adheres to brand guidelines, and allows for continued growth.

‍

That said, there are many best practices, both new and old, that many paid search experts follow. Implementing any of these could help improve your performance. At Pearmill, we developed our initial** structure/strategy by combining these best practices with learnings from our experiments. And it’s a methodology we’ve used to scale numerous search accounts over the years.

‍

Campaign Structure Breakdown

First, we have our campaign structure, consisting of one branded campaign and one non-branded campaign. Within the branded campaign, we typically have one ad group where we include both exact match and phrase match keywords. For our non-branded campaign, the amount of ad groups vary depending on keyword/product themes. 

‍

Let’s say that we’re setting up a search account. We’ll use a light bulb brand as an example, as a basic product that comes in many different forms. In this case, “lamp light bulbs” vs. “chandelier light bulbs” vs. “Christmas tree light bulbs” would be in three separate ad groups. This is so we can create better ads and landing pages catered to our audience and their purchase intent. Similar to the branded campaign, we’ll include both exact match and phrase match keywords in the same ad group. 

‍

Steering Google in the Right Direction

While Google continues to push broad match, we know from experience that this match type can bring in a lot of irrelevant traffic and waste spend. We would rather test into broad match keywords and do our due diligence – adding negatives to ensure we steer Google’s machine learning in the right direction. 

‍

Here is an example template of what our campaign structure would look like. Be sure to add your branded keywords as negatives to your non-branded campaign, so it doesn’t get the chance to trigger branded terms and potentially skew data. You want to be able to separate branded and non-branded data in order to gauge performance between the two more accurately, and to make optimizations accordingly. 

‍

Adding negatives between ad groups to prevent overlap in search traffic is also a good idea. For example, you wouldn’t want “Christmas tree light bulb” searches going to the “Chandelier” ad group, or else potential customers would be served an irrelevant ad. To avoid this, add “Christmas” as a negative keyword to the other ad groups. This ensures that Christmas-related terms only funnel through the “Christmas tree” ad group.

‍

Responsive Search Ads 

For each ad group, we have at least one Responsive Search Ad (RSA). In each RSA, we try to fill up as much space as possible to maximize the ad’s real estate, which means having fifteen headlines and four descriptions. To fill up more of the digital real estate and increase overall ad rank, it’s important to add these assets (formerly known as extensions). 

‍

We like to have at least four sitelinks, five callouts, and one structured snippet. Assuming there are multiple products at different price points, it’d be wise also to have a price asset along with a promotion asset for any sales you may be running. Image assets can help boost CTR and provide a visual complement to standard text ads, allowing customers to see the product before actually clicking through to the site.

‍

Dynamic Search Ads

In addition to our usual search campaigns, we may have a campaign that uses Dynamic Search Ads (DSAs). DSAs provide a simple solution for locating potential customers who are actively searching for products that you offer. These ads are particularly useful for brands with either a well-established website or an extensive inventory. By utilizing the content of your website, Google matches searches with terms and phrases frequently used on your site. 

‍

For example, if someone is searching “where to buy little light bulbs for Christmas trees,” then instead of targeting that specific keyword, we target the landing page with Christmas tree light bulb content. From here, Google automatically creates headlines for the ad while you can still create your own descriptions. DSAs can be used to complement your keyword-based campaigns and identify any missed relevant searches. 

‍

Remember to add your branded keywords and your non-branded exact match keywords as negatives to this Dynamic Search campaign so that it is forced to find new search terms that you can later target directly in your standard search campaigns.

‍

Account-Wide Negatives

Beyond these campaign-level negatives, we also want to have negatives that apply to the whole account. For example, if someone searches “UV light bulb,” but we don’t sell that product, Google’s intent-matching AI may still show our ad because it relates to our keywords. So we can add “UV light bulb” as an exact and phrase match negative on the account level.

‍

‍

Bidding Strategy

Now that we have our general campaign structure set, we can choose a bidding strategy and other settings that will help us have clear, concise data to work and experiment with. Regarding bidding strategies, for our branded campaign, we use manual CPC and keep bids very low in order to minimize costs while maintaining, or even improving, impression share. 

‍

For non-branded, we’ll likely start with Maximize Conversion Value to speed up the learning process and quickly gather data. Then after about two to four weeks, we’ll upgrade this to Target ROAS so that the system can play off a ROAS goal to keep in check. 

‍

Choosing the Right Networks

Along with the bidding strategy, there are a few more campaign settings such as the network you show ads on: Google Search, Search Partners, and Display Network. Google likes to push for advertisers to show on all of these networks to improve performance, but we know that results are often situational – some accounts react differently to certain settings, and when looking at the data, we see what’s really best for performance. We push our campaigns live while opted into Google Search only, and after some time we’ll then test into the other networks to see if they truly perform better. 

‍

‍

Location Options

In addition to showing on the right networks, we want to ensure that we are showing in the right geos. If our light bulb business only delivers within New York, then we don’t want to show ads to people searching for light bulbs in other states or countries. We’ll target New York and under “Location options” we’ll choose “Presence” instead of the default option “Presence or interest”. We do this to make sure we don’t run the risk of searchers clicking our ads in geos to which we don’t deliver. It’s another opportunity to reduce wasted spend.

‍

‍

Maintaining Creative Control

Lastly, we’ll opt out of “Automatically created assets” because we want control over how our ads are written. This is an important choice to make because Google creates these ads partly based on content from your site, which may not be suitable for an ad.

‍

‍

Conversion Events

Almost done – two final points that we ensure are taken care of: conversion events and auto-apply recommendations. Since we want to feed Google as much (good) data as possible, we like to include conversion events like “initiate check-out” and “add to cart.” 

‍

These will serve as secondary actions while the “purchase” event is our primary conversion goal, which our campaigns will optimize towards. Having these secondary events will allow for performance tracking throughout the different stages of the funnel and provide valuable insights about our user journey and possible optimizations for it. 

‍

We also set up our “purchase” event with a dynamic value and a default value of the average order (AOV) just in case there are any technical mishaps and someone’s order value doesn’t register in the system. 

‍

‍

Final Steps

Finally, we take a look at the auto-apply recommendations feature and disable all unnecessary options (which could be all of them). When turned on, these recommendations allow Google to make changes in your account that could result in compromised performance or budget misallocation. 

‍

As the platform continues to evolve and AI “simplifies” tasks, we seem to lose a bit of control. We want to maintain as much control as we can by not letting Google run wild with our clients’ money just because its machine learning says there will be an uptick in traffic next week. 

That’s it. That is our initial** structure/strategy for a search account. It serves as the basis for all future optimizations.

** This is a general template that we use to start many accounts but some may require a different setup – usually depending on client goals – and as we continue to experiment to try and improve performance, this structure may change over time.