Playing Moneyball with Google Ads

By Stephanie Courtney

Moneyball Now More Than Ever

The headlines across the UK, Europe, and the US are screaming about inflation and recession as the world heads towards an uncertain energy crisis in winter. While the pandemic felt apocalyptic, at least the house was warm while we were stuck in it.

With rising energy costs, people are going to have a hard time purchasing necessities. Gone are the days of a bit of expendable income from government subsidies during the Covid years.

What’s coming next might make 2021 look like a relaxing holiday.

Online retail is about to feel the pressure of scarce consumer spending, which many are not prepared for after the online boom of the last 3 years.

Jeff Bezos probably feels like the New York Yankees did in 2000 after winning 3 World Series in a row: rich and unstoppable. But just as there was a new kind of baseball brewing that would challenge the Yankees, there is also a new way of managing Google Ads that can give Amazon a run for their money.

The Moneyball Principle

If you’re British or European and you don’t know what Moneyball is, rent the movie. You won’t be disappointed.

For the sake of this article, the crux of what you need to know is in Michael Lewis’ preface to his book:

At the opening of the 2002 season, the richest team [in baseball], the New York Yankees, had a payroll of $126 million while the two poorest teams, the Oakland A’s and the Tampa Bay Devil Rays, had payrolls of less than a third of that, about $40 million….The raw disparities meant that only the rich teams could afford the best players. A poor team could only afford the maimed and the inept, and was almost certain to fail. Or so argued the people who ran baseball…The bottom of each division was littered with teams […] that had spent huge sums and failed spectacularly.

…On the other end of the spectrum was Oakland. For the past several years, working with either the lowest or next to lowest payroll in the game, the Oakland A’s had won more regular season games than any other team, except the Atlanta Braves. …How did the second poorest team in baseball, opposing ever greater mountains of cash, stand even the faintest chance of success, much less the ability to win more regular season games than all but one of the other twenty-nine teams?…

The answer was Billy Beane, the General Manager of the Oakland A’s, looked at the game and the data differently than any other team. He was able to find success in undervalued players who objectively played a hell of a lot better than they subjectively looked. Beane was able to be strategic because he looked at statistics that had performance value rather than falling into the trap of following stats at the heart of baseball myths that don’t determine what conventional wisdom says they mean.

The A’s came to Oakland in 1968, the year I was born across the bay in San Francisco. I grew up listening to Reggie Jackson play on the radio. I was at the Oakland Coliseum for a few of the games Ricky Henderson played, but until I cracked the book, I had never heard of Billy Beane, much less SABRmetrics.

SABRmetrics is simply the Society for American Baseball Research metrics defined as, ‘the search for objective knowledge about baseball.’

Perception Is A Mirage

I loved the book. Not because I’m a baseball fan. I’m not really. A bit, maybe. I loved Moneyball for what it reveals.

Moneyball codified something I’d discovered but couldn’t articulate while working in top, global advertising agencies like Group M/Mindshare, Carat/iProspect, and OMD: The big boys rigged systems to gain advantages based on money because they didn’t have the talent to compete by any other metric.

In digital advertising, big budgets do not generate the best ad campaigns, but they do buy the most views. Saturate digital media with ad impressions, and you can take credit for conversions whether the ad drove the purchase or not. Focus on the wrong KPI’s and present performance data in a certain way, and you can make the worst campaigns look successful.

While money can buy any media, it’s inferior and wasteful if the campaign doesn’t involve expertise and ingenuity. Conversely, talent, expertise and ingenuity can drive incredibly successful results, but big agencies revolt against this because quality takes effort, and effort takes time, and time costs them money rather than making them money.

Agencies are like the multi-million dollar players for the Yankees. Everyone wants to play with those guys, no matter how expensive they are (all of the big agencies I know and worked for charge fees based on a percentage of media spend – usually 15%) because there is a perception that they do it better than anyone else.

Except… those guys don’t actually win nearly as much as those of us that cost less, use real talent, and cultivated expertise. Big agencies buy awards and win contracts, but they don’t necessarily succeed in the digital advertising game.

It’s utterly counterintuitive – and so, as Moneyball revealed, is major league baseball. This concept is at the crux of the book’s subtitle: The Art of Winning an Unfair Game.

Levelling the Playing Field with Google Ads

Applying the principles of Moneyball to Google Ads is almost too obvious.

It’s so profoundly obvious, that Alphabet (Google’s parent company) could start a marketing campaign of their own, ‘Play Moneyball with Google Ads and Win.’

But Google may not want to admit that…because the principle of Moneyball is that Davids can do more than compete with Goliaths using Google Ads: Davids can win.

Since Google Ads makes the meat of their revenue nut from advertisers like Amazon, Proctor & Gamble, and L’Oréal among many of the other top advertisers in the world, it’s not in their best interest to encourage the little guys.

The Data of Winning an Unfair Game

By default, Amazon and the big agencies must automate all of their digital adverting. Scale alone requires complete automation.

Automation get things done. Automation is action. But it’s not quality.

Quality takes talent and manual effort to not only create engaging ads that drive conversions and revenue, but also to push efficiencies within the mechanics of campaigns that cut costs and increase value (ROAS – return on ad spend).

3 May: Revenue 103,861; Spend 9,173; ROAS 11.32

4 May: Revenue 102,449: Spend 7,954: ROAS 12.88

9 May: Revenue 103,019; Spend 9,063; ROAS 11.37

10 May: Revenue 104,177; Spend 7,736: ROAS 13.47

23 May: Revenue 100,916: Spend 10,902; ROAS 9.26

Quality shouldn’t come at the expense of automation. Automation should be used, but it must be appropriately harnessed to get the most out of it.

The golden triangle of Google Ads data are spend, revenue, and ROAS (Return on ad spend). ROAS is the equivalent of ‘on base percentage’ in baseball. ROAS not only tells you the rate at which your advertising spend is driving revenue, it also reveals how efficiently you’re spending.

Using real world examples, let me compare two recent days with similar media spends:

6 June 2022: Spend 12,260; Revenue 101,969; ROAS 8.32

19 July 2022: Spend 12,920; Revenue 89,400; ROAS 6.92

Not only is ROAS 20% higher in June, Revenue is 14% higher, and costs are 5% lower.

It gets better when I look at days in May 2022 when revenue topped 100,000/day:

3 May: Revenue 103,861; Spend 9,173; ROAS 11.32

4 May: Revenue 102,449: Spend 7,954: ROAS 12.88

9 May: Revenue 103,019; Spend 9,063; ROAS 11.37

10 May: Revenue 104,177; Spend 7,736: ROAS 13.47

23 May: Revenue 100,916: Spend 10,902; ROAS 9.26

The drop below 10/1 ROAS towards the end of May signifies when my client started to view media spend as the most important KPI.

Spend is to Google Ads as RBI’s (runs batted in) is to baseball. Winning baseball games doesn’t have as much to do with RBI’s as many players think, just as spend doesn’t have as much to do with driving revenue as most advertisers think.

Yes, you have to spend to drive revenue, just as you have to hit the ball to get home runs – but you don’t have to hit the ball to get on base (players get walked) just as you can make more revenue by spending less in Google Ads by achieving strong conversion rates.

Let’s dive deeper and look at CPC (cost-per-click) and conversion rates for those same days in May:

3 May: Interactions 16,024; CPC 0.61; Conversion rate 15.76%

4 May: Interactions 15,541: CPC 0.59: Conversion rate 15.27%

9 May: Interactions 16,183; CPC 0.58; Conversion rate 16.52%

10 May: Interactions 14,608; CPC 0.55: Conversion rate 17.51%

23 May: Interactions 17,321: CPC 0.68; Conversion rate 14.66%

It’s easy to jump to the conclusion that the drop in ROAS is due to the increase in CPC between 10 vs 23 May. The increase in cost is a factor, but not the most significant one.

The much more critical is issue is twofold. The 19% increase in interactions on the 23rd coupled with the 16% drop in conversion rates essentially drive the same amount of conversion volume for both days:

10 May: Interactions 14,608; CPC 0.55: Conversion rate 17.51%; Conversions 2,558

23 May: Interactions 17,321: CPC 0.68; Conversion rate 14.66%; Conversions 2,539

The fact that there are 19 less conversions on the 23rd after spending 3,166 more is counterintuitive – and it gets worse if I compare 23 May to 6 June conversion rates, and worse still if I add 19 July:

23 May: Interactions 17,321; CPC 0.68; Conversion rate 14.66%; Conversions 2,539

6 June 2022: Interactions 18,487; CPC 0.73; Conversion rate 13.70%; Conversions 2,532

19 July 2022; Interactions 17,085; CPC 0.82; Conversion rate 12.37%; Conversions 2,113

If we look at the hard numbers (volume rather than rates), the comparisons are striking:

3 May: Cost 9,173; Revenue 103,861; Conversions 2,526

4 May: Cost 7,954; Revenue 102,449; Conversions 2,434

9 May: Cost 9,063; Revenue 103,019; Conversions 2,673

10 May: Cost 7,736; Revenue 104,177; Conversions 2,558

23 May: Cost 10,902; Revenue 100,916; Conversions 2,539

6 June 2022: Cost 12,260; Revenue 101,969; Conversions 2,532

19 July 2022: Cost 12,920; Revenue 89,400; Conversions 2,113

How is it possible to spend more and make less revenue? How is it possible to drive fewer conversions from increased interactions? All things being equal shouldn’t conversion rates at least stay the same? Most importantly what changed from the beginning of May?

My client’s insistence that I spend more money without any planning, strategy, or time to build required automation. Automation takes little effort, but it comes at much higher costs and at much lower value. The difference between the first 4 weeks of May and after was due to automated campaigns.

Clearing the Way

The General Manager who brought me in 5 years ago, Neil Fitzpatrick, was let go at the end of May and the new GM and new marketing manager don’t understand Google Ads. To be fair, not many using Google Ads understand how it can work, and even fewer achieve palpable results. But Neil Fitzpatrick did. Neil is the Billy Beane of this story.

I met Neil when I was hired to manage the Google Ads account of one of the biggest retailers in Ireland. I won the pitch against 3 of the biggest agencies in Ireland. I was floored because I didn’t have a chance in hell of getting the contract – except that Neil saw and understood what I do. I was the equivalent of one of Moneyball’s undervalued players, who didn’t look like I could compete in the major leagues.

When he left that retailer to move country and industry, he took me with him. It’s the bumpiest transition I’ve ever made for reasons that could fill a novel. The salient point here is that Neil set impossible targets, gave me impossibly small budgets, and somehow inspired me to reach impossibly high revenue goals. But the most important thing he did was to get out of my way.

Neil gave me a budget, revenue targets and ROAS thresholds. Nothing else. What he knew best was that he didn’t know enough about Google Ads to manage anything more than those essential KPI’s. I was responsible for everything – ad copy, campaign types, determining which products to promote and which to leave off. Marketing managers, who knew less about PPC than he did, came and went and a lot of Neil’s genius was that he never required me to bow to their ignorance. 

As Google evolved, so did I, and what I discovered built not only The Alignment Collective but made my client’s website one of the top in their industry.

Neil understood that revenue volume has to gain value over time. This drove me to build and optimise the leanest campaigns imaginable and it drove millions in revenue at astonishing ROAS.

There were times when I would send in monthly reports and Neil and I would just marvel at the rate of return.

The new management don’t understand this. The new management and I are going our separate ways at the end of 2022.

When I started with this client, I had 10 years experience under my belt, a knack for building impeccable campaign structures, and a clear understanding of Google’s mechanics. By the time of Neil’s departure, this client’s annual revenue had increased 35% year-on-year since 2017 and that was almost entirely due to PPC. We had attained a lot of market share, and currently the only online retailer with higher share of voice is Amazon, and this client’s budget is a fraction of what Amazon spends.

Google Wins Whether You Play Moneyball or Not

No business can outspend Amazon – at least not in the Western world. The boon for Google Ads from Amazon merely being in the advertising mix is almost unquantifiable. Every £, $, or € spent by Amazon on Google Ads is at least matched (if not tripled or quadrupled) by thousands of competitors advertising the same products hoping to best Amazon in ad auctions that are almost rigged in Amazon’s favour.  

Almost.

Never forget that Google is in the business of making money – no matter what your Google Ad’s Manager says or how many free tools they offer.

But Google is still built in a way that can level the playing field. An ad for a mom and pop website can compete against an Amazon ad and win, with a click that costs less than what Amazon was willing to pay.

And yes, this kind of advertising can be done at scale. Neither mom and pop, nor mid-size website retailers will find this solution at big agencies. But there are a few small, independent Google Ad’s experts that can and do play Moneyball with Google Ads and win.

Amazon Is Not The Best…At Anything

Amazon isn’t merely outspending every other business in the world because it can. Amazon is outspending every other business in the world because it has to.

If you sell on Amazon, you know how utterly flawed the selling platform is – and having to pay ever increasing percentages of revenue in fees for the privilege of getting access to Amazon’s audience (while being restricted by Amazon’s idiotic rules) is truly frustrating.

But Amazon’s audience is undeniable – and one of the best ways to compete with Amazon is to snag potential customers that use Google (and to a significantly lesser extent Bing) to search for products before they land on Amazon.

If you buy from Amazon, think about why you buy from Amazon rather than other online retailers. In my experience, Amazon’s leverage comes from:

  • Unprecedented product inventory (a one stop shop for everything)
  • Consistent, relatively fast, shipping

And not much else.

The Amazon website is not well designed, easy to navigate, nor intuitive. It’s not a pleasurable experience, it is merely a utilitarian one. It saves time, but not money (contrary to popular belief). If you find products on Amazon, and then go directly to retailer’s websites (if you’re not buying directly from Amazon), products are usually up to 20% cheaper.

Competing with Amazon’s customer experience isn’t difficult. Creating a more enjoyable, efficient, cost conscious online shopping encounters takes more effort, but it is almost a no brainer because the bar is set so low.

Competing with Amazon’s reach in advertising is equally effortful, but the challenge is less about spending and more about ingenuity and quality. 

The most important thing about competing with Amazon is not to do anything like Amazon. They win because they pay to win. You can win because you deserve it.

The Myth of Big Agencies

Amazon spent $16.9 billion world wide on advertising in 2021 (that’s $1.9 million an hour). Driving net sales of $470 billion.

Amazon is the New York Yankees of online retail (much as that analogy diminishes Amazon, worth approximately $1.43 trillion, and elevates the Yankees, worth $6 billion, but you get my point).

Every online retail business competes with Amazon one way or another – just like every baseball team competes with the Yankees. Except if you’re competing with the Yankees, you’re one of only 906 exceptionally talented players in the world that is worthy of the major leagues.

There is no such discernment of retail worthiness to compete with Amazon. There are no minor leagues for online businesses that shield them from competing with the colossus. Amazon (and other competitors) challenges eCommerce websites irrespective of quality or performance – and the more Amazon destroys the little guys, the better their overall market dominance.

So the little and medium guys go to big agencies believing it’s the big agency that can compete with the big retailers. Except not only is this not true, it’s utterly detrimental. Going to big agencies that handle big brands compounds the problem; it does not solve it.

In my experience, the big agencies put all of their effort into their big clients who spend the most money (because at 15% of media spend, they have a vested interest in keeping big spenders happy and spending big). For above the line advertising, the little guys can benefit a bit from a big agency’s media buying power, but that paradigm does not exist for Google Ads.

On the digital marketing side, if the big agencies I worked at bothered to take on a small or mid-size client, we’d charge additional fees, and maybe – maybe – give them 5 hours or less per month of our attention. While the little guys budgets were low, the math of costs per hour where sometimes higher than the big boys – and the value of work they received was less than zero – because if the small or mid-size client was in competition with one of our bigger clients, we always made sure they couldn’t beat the bigger client.

Conversely, individual and independent Google Ads agents that don’t amass big client lists and don’t charge big fees, (the equivalent of the undervalued Moneyball players) can and do beat the pants off major league brands and major league agencies.

With Google Ads, there isn’t anything shielding Amazon, or any of the other Titans from the slings and arrows of small and medium online retailers that either hire in house gurus or find independent Google Ads agents who can and do chip away at Goliath’s market share.  And if you’re a David that plays smarter and better, you can beat Goliath, save money, win loyal, happy customers and stave off the doom of this recession.

Parting Tips

There are a million ways to run Google Ads, and I’m very sceptical about any accounts that are run the same way. Each account, industry, and audience is different. Audience behaviour and engagement matter most. Uniqueness, no matter how implausible, should be embraced when it drives revenue and ROAS.

The keys to Google’s slogan, ‘the right ad, at the right time,’ are:

  • impeccable campaign structures
  • a strong understanding of Google’s ad serving
  • focused targeting (using negative keywords/urls)
  • engaging ads (with price points/savings if possible)
  • precise landing pages

You can writing the best ads in the world, if your campaigns are sloppy/unstructured and your landing pages are not a natural progression of the ad, it won’t matter.

The biggest mistake I see is landing pages that are too restrictive. If the keyword used to trigger an add is ‘royal canin dog food,’ and your ad links to a Royal Canin gastrointestinal dog food page (because gastro is your top seller) then you’ve just paid for clicks for everyone looking for Royal Canin dog food, but you’ve lost the sales of everyone who wants to buy any other variety of the food. Based on the keyword, the landing page should include every king of Royal Canin dog food available on the website. And negative keywords should be used for any Royal Canin dog food not sold on the site.

Use automation to augment well-built manual campaigns, but don’t rely on automation. If you use Dynamic Search Ads, mine the search queries to find keywords that you can use to build static campaigns and ad groups. With strong ads, sitelinks, and price extensions, you can drive conversions and much higher ROAS than with DSA’s.

Negative keywords are imperative. Even with the new Pmax campaigns, you have to set up negatives for at least your brand terms or Pmax will cannibalize your brand campaigns usually at much lower ROAS. In order to set up negative for Pmax campaigns, you much contact your CSS and implement negatives with them. It’s a pain to do, but it’s worth the effort.

Negatives are even more important with the new ad serving of broad match terms – especially if you use Dynamic Search Ads. Google’s new broad match ad serving is too broad and leads to what I call chaff ads.

Chaff ads might seem innocuous enough, but they cause insipid and insidious little problems like increasing bounce and exit rates. I’ve seen countless clients go mad over bounce and exit rates, wasting lots and lots of resource to improve landing pages and spending budget on remarketing when the truth is chaff ads drove visits from users who were never going to buy.

Chaff ads also dilute conversion rates which muddy Google Ads machine learning.

Optimise shopping feeds – especially product titles. If you use .xml feeds, create a Google Sheet of optimised titles in a supplemental feed.

Test everything. Google’s best practices are rarely the best. Challenge every assumption and let the data lead you – just make sure you’re looking at the data objectively. It’s too easy to seek out data that proves your preconceived ideas rather than looking at data objectively – especially if the data is a bit baffling or counterintuitive. Strange data is usually where the veins of gold flow.

Distinction and excellence require hands on effort. Whether that means refining and optimising feeds before they are fed into the mechanisms of automation or whether that means building campaigns and crafting messaging by hand – or a combination of both – the way to compete with buying power is to subvert expediency with applied workmanship.