Are eCommerce brands too dependent on paid advertising?

This is an assertion that has been floating around in marketing and startup circles lately. In particular, we’re seeing a slew of new venture-backed eCommerce startups emerge with a glut of investor cash available to spend aggressively on advertising.

This was also the central premise of an article from Andrew Chen, who is a general partner at Andreessen Horowitz, a Silicon Valley venture capital firm. He argues that many companies die from their over-reliance and addiction to paid marketing channels.

He writes, “The key insight here is that paid marketing is tricky to grow, at scale, as the primary channel. It’s highly dependent on both external forces – competition and platform – as well as the leadership team’s psychology when things get unsustainable.”

Is paid marketing a dubious channel for eCommerce companies? This is an important question to ask.

Since we’ve helped hundreds of eCommerce brands over the last decade grow via paid and non-paid channels, we have a unique perspective to share with the eCommerce community.

This article answers this “overreliance” question by looking at a few core pieces of any eCommerce company’s marketing strategy including:

  • Attribution: Ensuring you have accurate measures of cost of acquisition by channel
  • Scale effects: Our thoughts on how paid channels diminish or don’t in efficacy with scale
  • Saturation: Most brands are at risk of “saturating” a market with paid ads

In short, while we agree with Andrew on a couple of smaller points, based on our experience,  we feel that most eCommerce brands shouldn’t worry about an overreliance on paid marketing as long as attribution is properly accounted for and return on ad spend (ROAS) is monitored.

Let’s get into the details.

Understanding True Customer Acquisition Costs

A key component of this “overreliance on paid marketing” argument is that a company needs to have an accurate measure of each channel’s Customer Acquisition Cost (CAC) and that simply monitoring “blended” Customer Acquisition Costs (CAC) is dangerous, as it can lead companies to waste thousands — if not millions of dollars — on unprofitable ad spend.

We agree.

While Blended CAC means slightly different things to different people, in general, it refers to just tracking a single number for acquisition cost that is total sales divided by total marketing costs.

Obviously, this leads to the danger of profitable channels obscuring the existence of unprofitable channels.

For example, if a company was losing money on Facebook ads in a given month, but had their product picked up by Good Morning America and got a bunch of sales that way, their overall company CAC could look good. It would hide the fact that Facebook ads were not profitable (or not as profitable) as they think.

But, this problem isn’t an issue as long as you’re looking at CAC by channel. And in our experience, the vast majority of eCommerce companies today do this already.

For our clients, we look at CAC, and more importantly, return on ad spend (ROAS) at a far more granular level than just channel. We look at CAC and ROAS:

  • At the platform level (Adwords, Google shopping, Facebook, Instagram)
  • At the campaign level
  • At the ad group level
  • For specific keywords, products, terms, creative, and more

So we have an extremely detailed understanding of ROAS along all paid acquisition programs down to specific ads and attributes.

We recommend all eCommerce companies think of CAC and ROAS at this granular of a level.

Because this depth of understanding lets you do things like lower cost per acquisition (CPA) from $47 to $9 in a Google Shopping program and develop best practices for how to compose adwords ads for the highest ROAS.

So, while blended CAC isn’t an issue, there is another attribution problem that many eCommerce brands fall victim to that we’ll explore in the next section.  

Analyzing Last-Click v. Multi-Touch Attribution Models

One mistake we do see eCommerce companies make, however, is just looking at last-click attribution in Google Analytics. This is the default attribution for Google Analytics.

In our view, this can really handicap a company’s understanding of

  • The true ROI of paid ads
  • The role that paid ads take in a customer’s entire journey with your brand.
Last Click Attribution Model example

Last-click attribution makes it seem like your customer acquisition is linear, takes place on a single device, and then gives all the credit to the last touch point. A customer sees an ad. Then, clicks the link and buys from there. If they buy immediately, the ad gets credit. If they later come in via the homepage or organic search and buy, then the ad would get no credit under the last click attribution model.

Obviously, this framework is extremely limited.

In today’s market, most customers will have multiple touchpoints with your brand before they buy. The average customer may click on your ad, read a few of your tweets, check out some product reviews and then Google their way to a blog post (Note: We’ve shared multiple case studies on eCommerce blogging and content such as buying guides helping with sales) – all before they ever make a purchase. When you look at it from the lens of last-click attribution, the blog post will get credited with the sale. The reality is your paid advertising, social media and content all assisted with that sale.

A flowchart showing how sales work. Generally, consumers don't immediately purchase during the same visit.

One way to improve your view of the role these multiple touchpoints play in “assisting” with the ultimate conversion event is by analyzing sales via one of several multi-touch attribution models. Multi-touch attribution is not perfect, as it still doesn’t account for cross-device touches (i.e. if someone click an ad on their phone then later buys on their computer, you won’t know it was the same person and that the ad contributed to this sale), but it’s a clear step up from last click attribution.

For example you can use a few GA reports to look into this. For example, using their built in “model comparison tool” you can evaluate sales via last interactions (simple last click attribution) or other models, such as “position based”, where first and last are given more weight and the middle interactions less:

Google Analytics attribution model sample

Here is Google’s explanation of their different attribution models.

Scale Effects Don’t Have To Work Against You With Paid Marketing

Even after properly calculating acquisition costs, understanding attribution, and ensuring that each paid platform, ad campaign, or even each ad is profitable and meets your company’s target ROAs, the “over-reliance on paid marketing” theory argues that eventually, as you try to scale paid channels, your ad costs will grow.

In the Andrew Chen article we cited at the beginning, he argues:

“Saturation is also a thing. As you buy up your core demographic, the extra volume comes from non-core, who are less responsive.”

We think this worry, that you may “saturate” your market, should not be a big concern for the vast majority of eCommerce companies.

Here’s why.

First, many eCommerce companies are selling products that have a steady and consistent demand from new customers. For products like this, you can use paid marketing to consistently get in front of new waves of consumers looking to purchase your products.

For example, let’s assume you manufacture and sell furniture. Furniture has been in consistent demand for centuries. You will always have a steady influx of new potential customers looking to buy furniture. So we think this view that there is some “core demographic” of customers who, once you show them enough ads, will tire of them, and you have to move on to less profitable demographics is faulty because:

  • There are paid channels like search, where you can serve ads on extremely high intent keywords: “Black 3 seater leather sofa”
  • The people searching for that next month are likely different from those searching this month, thus you aren’t wearing out a single group from over exposure

Second, as we mentioned above, there are so many digital touchpoints today (versus when Dropbox was testing paid ads 10 years ago, as in Andrew Chen’s article), that viewing paid marketing simply as a direct, last click acquisition channel and nothing more is not sound.

In many industries, consumers need to see and build trust with brands before deciding to purchase. They want to compare brands. They want to educate themselves about products. They want to read reviews and see social proof. All of these touchpoints can be aided with paid marketing.

Thus, we feel that a holistic paid marketing strategy for any eCommerce company will involve getting in front of the consumers in all of these situations.

As you spend more and more, are there diminishing returns from a ROAS perspective? No doubt. But will most brands really saturate all paid channels and not be able to find as profitable channels to spend more capital? We think this is highly unlikely. Of course anything is possible.

If you truly have enormous amounts of investor capital waiting to be spent, perhaps you’ll reach saturation. But even after working with hundreds of eCommerce companies for over a decade, we haven’t run into it.

Conclusion

We argue that for the vast majority of eCommerce brands, you are unlikely to hit 100% saturation levels. Instead of pondering it, focus on improving ROAS. This includes attribution modeling and understanding all of the assisted conversion touchpoints that your customers have before they buy from you.

Want a custom in-depth assessment of your paid marketing performance? Contact us to get started.