One of my favorite questions for new professional acquaintances is, Whats your favorite innovation from the past six months, and why?

I like this question for a few reasons:

Automation almost always comes up. PPC marketers are spoiled for choice on ways to automate their workflows, such as:

This article will not pass judgment on which automation options you might choose to opt into. Rather, it will outline the mechanics of each, as well as how to make them work.

Bidding arguably is one of the few things that should always be automated.

How advertisers automate is where theres room for different paths to profit:

If you opt for native bidding, its vital that you trust your conversion tracking. Ad networks use conversions as the main guiding light to know whether to invest or pull back.

While its true that Max Clicks and Target Impression Share dont need conversion tracking data to function, its still vital that youre feeding in the right data.

A common mistake people make is including too many conversion actions as primary. This usually results in double counting or including steps on the journey that arent worth feeding into the algorithm.

Another common mistake is setting a target that isnt attainable. If the target cost per acquisition (TCPA) or target return on ad spend (TROAS) dont match the budget and auction prices, youll likely underspend or flood your campaign with bad leads.

Rules and scripts are helpful when you have enough data to know what your bids should be. They typically need larger budgets to help compensate for being on manual bidding (and forfeiting the native bid signals).

The most common signals are:

While these do require healthy data, the biggest pitfall for rules/scripts is the human element.

If the input is incorrect, it might hurt the campaign.

This used to be my favorite way to automate bids: set a conservative manual bid and go aggressive on the bid adjustments.

Bid adjustments could be used to direct budget away from, or towards, desired prospects and give the user the most control.

However, with the improvements to native bid adjustments, its harder to justify using these. If you are planning on running a purely manual campaign, these are a must.

If you trust your conversion tracking, theres no reason not to use native bidding.

Native bidding pulls in signals advertisers arent able to access and usually outperforms human-governed automation.

Max Conversions with a TCPA goal and Max Conversion Value without a goal beat manual biddings 6.8% conversion rate.

However, if you arent able to put in accurate conversion value data or set realistic TCPA goals, going manual with rules, scripts, or bid adjustments will serve better.

When asked, many will say they prefer human creative over automatically created text, display, and video ads.

The human mind is supposed to be better at coming up with the right content for the right audience.

However, what may go unnoticed is that the lions share of auto-generated ads come from human creative. i.e., the text, tone, and images will be borrowed from the advertisers site, existing ads, or social channels.

Creative falls into these categories:

Responsive ads and PMax ads all fundamentally follow the same objective: Take the content the advertiser provides and find the best combination for those assets.

Marketers can choose whether to allow full learning or to help it along by pinning creative to desired locations.

It can be tempting to pin all assets to recreate Expanded Text Ads. However, as this Optmyzr study shows, youll get a higher conversion rate by allowing for some freedom in the learning.

Visual creative (static display images and videos) are a bit tougher.

Brand style guidelines can make auto-generated ads a no-go. Be sure all creatives work with the templates, and if they dont, consider uploading a fully contained ad.

Dynamic Search Ads (DSAs) are powerful; however, much of their utility has been absorbed into PMax. Part of their power is leveraging the well-SEOed nature of a site to help Google know which landing pages relate to a query.

Headlines are derived from the language on the landing page.

Currently, its not a question of when almost all creative has automation baked into it, to a certain extent.

The main choices to make are whether to fully delegate creative control or partner with the machine.

Assets (formerly known as extensions) should almost never be automated. This is because you want to use them to their fullest: highlighting high-value services, and amplifying ad messages.

Audiences are the beating heart of PPC campaigns and the main lever advertisers have at their disposal.

However, the privacy-first web has restricted which audiences we can use, and even saw the depreciation of some of them.

Marketers access audiences through the following:

Choosing to opt into any of the automated audiences (i.e., any that a human doesnt actively choose) depends entirely on how much you trust your conversion tracking.

Its also important that your creative matches the audience youre targeting. As a general rule of thumb, opting into automated audiences means manually excluding audiences you dont want (provided there are no restrictions).

Manually choosing audiences (particularly opting into audience signals for PMax) invites room for human error. This is why its important to use data to inform your choices.

Consider building custom search intent audiences based on your top converting search terms (as opposed to the keywords themselves).

Unless youre in a restricted industry, it almost always makes sense to have some manual influence on the audiences. This influence can be exclusions or outright targets.

However, as broad match, smart bidding, and PMax have evolved, its worth testing the baked-in audiences.

At best, they thrive, and youre able to benefit from leaning into the ad platforms toys.

At worst, you have data to harvest and can be proactive in future exclusions and ad creative choices.

This is one of the biggest areas for contention with marketers. Marketing channels are a strategic choice, and delegating this level of strategy can cause understandable uneasiness.

However, depending on how the human interacts with the automation, it can still be a net positive experience.

Automating ad channels boils down to the following:

Sometimes PMax can outperform campaigns with more controls in place. This is usually due to the following being true:

Many marketers default to search first, and not all prospects want to consume information that way.

Additionally, by the time the person is searching for the thing, they usually have a shortlist of brands theyre interested in.

PMax enables brands to access visual placements and top-of-funnel interactions without carving out a separate budget for it.

It is important to have account-wide negative keywords and placement exclusions for any automated channel campaigns.

Manual campaigns can be effective they just tend to require advertisers to pay a premium for each placement and are limited to the channels the advertiser opts into.

The biggest deciding factors will be budget and conversion data.

If you trust your conversions (and are able to include conversion values), automated marketing makes a ton of sense. The human marketer puts safeguards in place via negative keywords, audiences, and placements.

If you dont trust your conversion tracking and budgets are restricted, youll want to be specific about where marketing dollars are invested.

Its important to note that image extensions (available on both Google and Microsoft) are a valid answer to multi-channel campaigns that need more control.

If you are under strict brand standards and/or have to report on all channel ROI, that might be a more feasible solution for you.

Automation is ultimately a net positive for digital marketers.

How much folks engage with it depends on which tasks are core competencies and which are struggles.

The more a brand trusts its data, the easier it is to lean into automation.

More resources:

Featured Image: /Shutterstock

Read the original post:

PPC Automation: When & How To Use It - Search Engine Journal