uavsdirectory.com

Module 4: Google Ads (Search), Metrics and Performance

In Module 3 you learned a framework for creating high-performance Google Ads campaigns. In Module 4 we’re going to discuss what to do after those campaigns start generating traffic.

In this module I’m going to try to convince you of two seemingly contradictory ideas. First, that understanding advertising account metrics and performance is absolutely critical for your success with paid advertising.

And second, that metrics and performance data are generally worthless.

How can those two things be true? Let’s dive in and I’ll explain. Along the way you’re going to learn how to use metrics better than 90% of the people I’ve met who manage advertising accounts professionally.

Before we do that, here’s a friendly reminder: it is highly likely your initial results will be poor, especially if this was your first attempt using Google Ads (or your first attempt with a very different methodology for Google Ads). I want to reiterate that poor performance initially is almost always a fact of life in paid advertising — it’s what we do next that matters.

This module will give you a road map to understand how to make those decisions.

Let me define the terms I’ve used in the title of this module so we have a clear, shared understanding.

Performance is the measurement of how your ads are contributing to the primary goal of your specific business system.

Metrics are the numerical values we use to determine performance and manage our advertising account.

In Day 6 of the free paid traffic training (“Know Your Numbers”), I mentioned two types of metrics — business management metrics and account management metrics. It’s important that we understand the difference.

Business management metrics express high-level outcomes relevant to the financial management of your business. There are four that I find most valuable — cost-per-acquisition (CPA), average order value (AOV), lifetime value (LTV), and Return on Ad Spend (ROAS).

We’ve discussed these numbers in a previous module. As a refresher, CPA tells us how much it costs to acquire a paying customer. AOV tells us how much that customer spends on the first sale. And LTV calculates how much a customer spends over the lifetime of the relationship. These are the primary metrics of direct response marketing and they’re conceptually useful beyond direct response as well.

ROAS is often used in e-commerce (although it can be used elsewhere as well, including any offers where there’s an expectation of profit on the first sale). The calculation for ROAS is gross revenue / cost.

A challenge with using ROAS is finding the appropriate level to calculate. Total revenue / total advertising cost will tell you in very general terms if your paid advertising is ROI-positive. But, at that level of analysis, it gives very little insight into actions to take.

I prefer to calculate ROAS in Google Ads at the keyword phrase level whenever possible. At that level we know which phrases are generating the most — and least — revenue relative to their cost and can act accordingly.

Let’s move on to account management metrics. This is where most of the problems with using metrics can be found.

I review account management metrics in Google Ads, and support those with metrics in Google Analytics.

These are the metrics I most commonly review in Google Ads:

These are the metrics I most commonly review in Google Analytics:

At first glance, that looks like a lot of data to consume and understand. Don’t worry — one of the reasons we’re starting with only one search campaign with one search phrase is so we can learn how to evaluate performance effectively without feeling overwhelmed.

This is an important skill that you will practice repeatedly until it becomes second-nature.

The most common question I hear when talking about metrics is “what’s a good (fill in the blank)?”

  • What’s a good Quality Score?
  • What’s a good Clickthrough Rate?
  • What’s a good cost-per-lead?

On and on it goes…

Let me be very clear about something — those questions are meaningless. How much money does your business generate because of your Quality Score? (Answer: none.) Where do you deposit your check for a high clickthrough rate? (Answer: you’re not paid for a high clickthrough rate).

But surely, you ask, cost-per-lead must be meaningful, right? On its own, cost-per-lead is meaningless too…

The best way to explain why metrics, on their own, are meaningless is to show you a real example.

There are two critical insights to understand. First, we use metrics in service to performance — we do not manage to metrics for their own sake. Managing to metrics is among the most common mistakes I’ve seen in my career.

Second, we assess performance within a broader, system-level context. Metrics only matter relative to what we are ultimately optimizing our business for.

The best way to internalize and use these two insights is to translate your metrics into the story they’re telling. I do this regularly for myself and my clients. Story requires context, and that forces us to see and articulate the broader view.

Let’s consider some of the stories I could tell the client I referenced in the video.

Story #1 — They spent $40,343 on a Google Ads campaign that produced forty customers at an average cost of $1,009/customer. (That’s true, but not meaningful or useful.)

Story #2 — They spent $40,343 on a Google Ads campaign that produced forty customers at an average cost of $1,009/customer, which is $209/customer ($8,360 total) more than their target CPA of $800. (That’s also true, and it’s in a larger performance-focused context, but it’s not meaningful or useful.)

What’s the logical conclusion from each of these narratives? This Google Ads campaign doesn’t work. Is that true? Absolutely not. Parts of the campaign are wildly successful. Turning off the campaign would be foolish.

Here’s how I would write the narrative to this client.

“From March 1 — April 15, 2020 you spent $40,343 on a Google Ads campaign that produced forty customers at an average cost of $1,009/customer. That was $209/customer ($8,360 total) more than your target CPA of $800.

At the campaign level, this was not a success. However, if we look deeper it’s clear to me that we can make it successful.

Within that campaign were three keyword phrases that generated customers within the target CPA range. Those three phrases created eighteen customers at costs of $147 (11 customers), $559 (5 customers), and $431 (2 customers). There is no additional traffic available for those keyword phrases. Because all three were within the target CPA, I recommend leaving those on.

Two other keyword phrases have the potential to reach the $800 threshold. One produced fifteen customers at $938/customer, and the other produced one customer for $1,060. There is not enough potential for the keyword phrase that produced the $1,060 customer to justify spending time optimizing that. I recommend spending $1,500 total to see if there’s another conversion. If there is, we’ll continue advertising for that phrase. If not, we’ll turn it off.

The keyword phrase that produced fifteen customers @ $938/customer has significant potential. It is a large audience and we only reached 86% of that audience in this time frame. Optimizing that keyword phrase has the potential to produce a significant number of customers per month and that’s where we’ll focus our energy and attention.”

Putting those numbers into context makes them meaningful and useful. With the client’s approval, I would then dig deeper into the data to determine how to optimize cost per conversion from $938 to $800 (or less).

To do that, I would first look at the Search Terms report in Google Ads (formerly the Search Query Report). That report shows the actual search terms that triggered ads. I would look for the search phrases that produced conversions, and then look for opportunities to eliminate low quality searches with negative keywords.

For example, let’s assume I looked at the Search Terms report and found the following (very simplistic) data:

  • Buy red widgets — 100 clicks, 8 conversions
  • Buy red widgets online — 50 clicks, 4 conversions
  • Buy red widgets online free shipping — 30 clicks, 3 conversions
  • But cheap red widgets online — 100 clicks, 0 conversions

What would I do? Add cheap as a negative keyword (-cheap) — it’s generating a lot of clicks (which means a lot of expense), but no conversions.

Sometimes a keyword phrase can become instantly profitable simply by adding a few negative keywords. Not always, of course, but you would be surprised how often a simple change like this can work.

If I couldn’t find what I was looking for in the Search Terms report, I would look at the landing page next. Because this example keyword phrase has so much potential to generate results, I would be inclined to create a landing page specific to that search query so I could speak to the exact needs of that audience.

What’s next?

The only way to really understand metrics is to roll up your sleeves and start interpreting your own data. To do that, you’ll need to understand context so the data you see has meaning. And, in order to accurately describe context, you’ll need to determine what you’re optimizing for, and how the system you’ve produced does that.

As I’m sure you’ve realized, my method for evaluating metrics is designed specifically to be an upstream intervention. In order to use the method effectively, you need to take a lot of positive actions (like identifying the goal of your system, determining what makes the boat go faster, knowing — or knowing how to calculate — your high level metrics, knowing the structure of your business system, and seeing how paid advertising performs within this larger context).

Then, and only then, can you write the story your data is telling (and adjust your actions accordingly).

What if you’re just getting started and you don’t know all of the details of this larger context? Then you need to decide what you do know, or what you can know.

For example, if you want to build an audience for a co-created offer, but you’re not sure if your ideas will resonate with that audience, you don’t need to spend a fortune on paid traffic. Instead, decide that the goal for paid traffic initially is to generate 100 leads, that you’re going to email each of those leads personally asking three questions, and you want at least 40% of those leads to respond to your emails in order to proceed.

In this example, you’re optimizing your system for producing high-quality leads (as measured by the response rate to your email). When you conduct an analysis similar to my earlier screencast, you can place all of your data within your specific context.

It might sound something like this:

“I created a Google Ads campaign with two keyword phrases that I thought might generate high-quality leads for a co-created offer. Keyword phrase #1 generated 60 leads @ $2.50/lead, and keyword phrase #2 generated 40 leads at $3.50/lead.

In order to proceed, I wanted 40% of the leads to respond to my personal email.

Thirty of the leads from keyword phrase #1 responded to my email and twenty of the leads from keyword phrase #2 responded to my email.

In total, I spent $290 to generate 100 leads and fifty responses, exceeding my goal for the test.”

Then, you would determine your next actions for your specific situation, continually iterating your way toward success one test at a time.

That’s a very focused way to use paid traffic to get meaningful results without fuzzy goals.

NEXT: TTE Q&A Call #2, Modules 3-4