Marketing Insights
Marin Software's Official Blog

Built to Last: What Automated Bidding Can Do for You

Built to Last: What Automated Bidding Can Do for You

By   August 29th, 2019

When I first started working in digital marketing, part of my training process was learning about best practices, campaign structure, and all kinds of additional tools provided by publishers and third party vendors to drive success for brands.

As discussed in my previous posts, it’s important to know your brand and your customers, and to understand what makes them tick (and click) to convert on your website.

One way is to continuously run tests on your activity and learn what works and what doesn’t. Or, you can switch on and off tools like audiences and bid adjustments, explore new campaign types and channels, etc.

This article is dedicated to bidding—automated bidding, to be more precise. However, it would be wrong to look at this topic separately from everything else that’s happening in your accounts.

Much like in your home, wi-fi wouldn’t work with no electricity. Or, how can you enter the living room without a front door?

Let’s walk through our house together from room to room to see how every component makes our environment cozy and inviting to our guests—in other words, our customers who we’re driving to convert and come back for more.

Let’s start with the hallway (a.k.a., data source)

When thinking about your account optimization and bidding (whether it’s manual or automated), it’s important to identify your trustworthy data.

Some advertisers have developed their own tracking and attribution systems in-house. These systems provide them with first party data, and all the insights they opted into their product to track user journeys and interaction paths with the brand.

Another option is to pick a third party vendor or publisher tools to execute tracking for you. Here, it’s important to review the packages vendors offer and understand which ones suit your business needs the most.

Remember, the digital world is changing every day. Online customer behaviors simply aren’t the same as they were five to 10 years ago.

For instance, social networks now have more influence on a consumer’s decision making process. When someone’s in their exploration phase, reviews and feedback are important. This assigns higher revenue share from sales to publishers like Facebook, Instagram, Pinterest, etc.

The question here: as an advertiser, are you taking cross-channel user interactions into account, and is the value you assign to your social media or affiliate advertisement fair?

There are various vendors that consider these factors in their attribution model and share the conversion data between the channels fairly—whether it’s a click from Google or impression from Facebook that contributed to a conversion.

Marin Software has developed an attribution model called TruePath, which helps advertisers see the impact of other channels on their search campaigns and vice-versa. This allows you to better understand the value of campaigns that don’t necessarily convert on their own (e.g., prospecting), but play a key role in the user journey.

Sugar, spice and everything nice (a.k.a., audience
kitchen prep)

In past posts, we discussed the importance of understanding your audiences, how they affect performance, and what it takes to bring home that extra conversion.

When you’ve identified a list of audiences that convert better than others, you can add these and their similar audiences to specific or all campaigns. As these are best performers, you can add specific bid adjustments to these.

With one of my clients, we created audience rules to cover their top product categories and certain pages—like What’s New? After running these audiences for a month, we saw that after visiting the What’s New? page, people were 32% more likely to convert on the website, even if they purchased long-existing stock. The next step was obvious: increase the bids on audiences that visited the page and reap the results.

There are tools that offer automated bid adjustments for audiences tagged across campaigns. For instance, the Google eCPC model takes audience behavior into account (Google conversion pixel data only) and adjusts bids for you based on audience performance in the campaign/group.

Marin Software also measures audience performance when it comes to automated bidding. Here, however, the app can use your preferred source of truth. You can select a list of conversion types that are important when it comes to bid adjustments and performance and the app will do the rest—calculate and execute.

At Marin, we’ve also come up with the concept of search and social intent. Here, we help our clients to reach social prospects and leads in search engines and vice-versa.

For example, we ran a test on social campaigns using search data. The result: CPA for these campaigns was 65% lower in comparison to traditional social campaigns, while conversion volume spiked by 388%. The client also saw a 100% increase in their daily appointment requests.

The living room (a.k.a., device performance
and adjustments)

According to many marketing sources, every year is The Year of Mobile. 🙂

And yes, every year, we witness how mobile search share is growing, as mobile devices become more versatile and occupy every part of our lives. This may or may not be the case for you, but I feel lost without my phone, as it’s such a huge part of my life—all of my contact information is there, not to mention the easy ability to call for food delivery or an Uber in just a couple of swipes/taps.

And yet, when it comes to conversion rate and return on investment, desktop keeps showing higher, better results.

Why is that?

SmartInsights and many more research companies compiled research on this topic to prove that even though most of the initial searches on the product/service/ideas are coming from mobile, conversions are more likely to happen on desktop.

A key takeaway here is that it’s important to review device performance not only against each other, but also from a cross-device perspective.

Marin has created a special solution for this—you can set separate targets to devices. Then, based on the targets and their performance, our solution adjusts bids to meet the desired goal.

 

Home office (a.k.a., additional rules in Marin bidding)

Marin bidding equips you with a list of additional rules to augment your usual strategy. You can apply these on top of your target. These include:

  • Bid cap
  • Bid floor
  • Minimum daily bid change
  • Match type boost
  • Publisher boost

Is your new promotion about to start outside of your working hours? You can set up your schedule to boost bids for a fixed time period.

Or, is your finance team keen to stay on top of the maximum bid values? Not a problem—just set up a bid cap and Marin won’t assign bids above your desired value.

The app uses these rules as advanced triggers to keep your performance in check. It also provides the calculation used for every object on bidding to show what was used in the process, and why.

When it comes to adjusting your target/strategy, the app makes everything very transparent and easy to understand.

Behind closed doors—what else can we do with our campaigns?

Device, audience, geo-targeting, time-of-day and day-of-week—these are the first adjustments that come to mind when reviewing a bidding strategy. But can we do more?

You bet! Our Marin analytics and product teams have worked hard to bring extra zest to your mix-and-match bidding approach.

For instance, do any of these scenarios sound familiar?

  • Your top-selling shoes are out of stock but your campaign is live
  • Impression share has dropped but you’re trying to stay ahead of competitors
  • A list of objects needs an extra push because of a new product launch or seasonality

Who you gonna call? In this case, not Ghostbusters—but, feel free to pick up a phone or email account and contact your Marin team.

Marin’s Dynamic Actions is your secret weapon when it comes to bid adjustments based on your unique situation.

With Dynamic Actions, you can tag objects in Marin with certain labels, whether it’s product stock or a certain performance metric. Marin will then recognize this tag, and based on your preferences, bid down or up for selected objects. Or, you can even stop the bidding altogether if you run out of stock.

This comes in handy when you have additional metrics you want to consider for bid optimization.

Making the place comfy (a.k.a., the importance of
bidding target)

Whether it’s revenue or exposure, every organization works towards certain goals, and it’s important to hit them to grow your business.

With automated bidding, your main responsibility is to identify the desired target and to experiment with adjustments and data points. This maximizes the benefits of machine learning for an optimal bid calculation.

When you’re assigning a target to a group of objects, bear their current performance in mind. For example, is the target too aggressive? If so, perhaps it’s best to start +/-20% of the current performance and adjust the target every couple of weeks. This allows the final bid to change smoothly over the given time period, and avoids performance spikes.

If you use the bid adjustments above wisely, they’ll bring great value to your advertising programs. To test them out, start with one to measure the impact, and then add or remove as you go.

As we mentioned in our audiences article, it’s important to have statistical proof that a certain element is bringing your performance to the next level before you add it to your initial target. Also, remember the importance of your budget—sometimes, a higher bidding budget for your test can bring in incremental revenue at a low extra cost.

For instance, during a bidding test with one of our clients, Marin managed to increase lead volume by 600%, while the average CPL in their account dropped by 20%. The team also saved 25% of their day-to-day time by letting the app handle bid optimization.

If you have a little extra budget on hand for the test, it can lead to excellent results.

Last but not least—cleaning house (a.k.a., bidding can’t work on its own)

Some advertisers ask us: why is it important to continue optimizing campaigns, when automated bidding is switched on?

Even the most modern and automated house—where Alexa and Google Home answer the door and adjust the lights—needs a human touch.

Automated bidding can indeed help you save tons of time when it comes to keeping tabs on performance and updating bids. However, we have to remember the world around us is constantly changing. Therefore, digital campaigns continually need fresh ad copy and images. And, with our ever-expanding vocabulary and trend shifts, search engines see +15% new search queries every year.

So, dust the surfaces, make them shine, and be able to show off your best performance to date.

P.S. look at our beautiful garden (a.k.a., bidding
test example)

Hope you’re still with us on our home maintenance metaphors!

Like any other top performance marketing company, at Marin we continuously run bidding tests with our clients. Let’s take a look at a test we performed against Smart Bidding, where the goal was to maximize leads to a set target cost per lead/acquisition (tCPL).

Within Marin’s bidding folders, the team activated portfolio bidding functionality. This functionality allows the app to use Bayesian data blending to help the algorithm identify the best bid for the given keywords. On top of this, automated mobile bid adjustments accommodate performance differences across devices. They also ensure placement on the first page of the SERP, and a high position and impression share at competitive levels.

In Google Ads, the client set the Smart Bidding tCPA model for a set of campaigns they used in the test.

During the test, the client didn’t create any new creatives or updated keywords. This kept the data as fair as possible across the campaign buckets that were in competition.

A month after the test started, the team compared results to the time period before the test. We ran the test during a low seasonality timeframe to avoid any external factors that might interfere with the results.

The post-test results were stunningwhile running pure automated bidding activity in the account, both solutions showed an increase in conversion volumes MoM +51% with Marin’ bidding solution and +44% with Google Smart Bidding. The CPL also dropped, meaning that not only did the client gain incremental conversions from the automated bidding, but they also did so at a lower cost.

CPL in Marin dropped -47% MoM and -28% in Google Smart Bidding.

When all is said and done, when it comes to bidding, it’s important to evaluate all tools an advertiser can and should use. Once you identify areas of performance, you can use them as a layer to your bidding strategy. Automated tools help with the process of calculation and execution for hundreds or even thousands of keywords—with no time or effort from your team.

So, agree on the strategy, set it up, sit back and relax, and let the machine deliver the best bids possible.

Now that you have the perfect automated bidding home—go take that much-deserved vacation!