Here’s what I learned this week

We had a few clients churn last week.

Some I expected, some I didn’t.

But upon feedback, one thing is super clear -

There’s a huge gap in the understanding and application of data in today’s ecommerce market.

Relevant post:

When questioned the overwhelming response pointed to “list growth” thinking that software is the answer to magically unlocking this.

There’s no magical list growth with just software, you can only control

1. Offer

2. Timing

3. Audience

That’s it.

The other response I got is that “they learned what they needed to know” unless the subscription to conversion rate gets to 100% across the board, then I can 100% guarantee you, no one has learned enough to build out the frameworks and track them towards that level of improvement.

Sometimes my frustration with the overall industry bubbles up to the top and I start to question behaviors when they no longer match with basic logic.

I’m not mad, I’m just disappointed.

So it’s fitting that I drop today’s publication.

I’m going to tell you exactly how we run Facebook Ads with data, it’s an indepth view of what we do to prevent failure points without going nuts testing zillions of creative.

Facebook still works, it just requires a shift in strategy moving forward

I've been seeing all the posts on reddit, twitter, and linkedin about ad performance on Facebook, saying that Facebook broke things.

People freaking out about bad days and good days.

I think when everyone is doing the same things and taking the same approaches eventually saturation sets in and we need to shift our overall strategy to incorporate data that Facebook no longer has access to.

We stopped paying attention to ROAS a long time ago.

ROAS has never told you if someone purchasing is a good purchaser.

With funnels people using to convert people with big sales and other intro offers, sometimes going this far contributes to a lot of people that never purchase again.

Those are what I like to call long term business killers.

When looking through the data, we realized there were other elements of the audience that were more correlative to the success of the messaging of the ad to drive high quality audience and ultimately a sale outside of that 7 day period.

There was too much variability where at small ad spends a few large orders would completely skew the results for a day.

On occasion an ad set spending $400 a day would have one or two $200 dollar orders come through and massively skew the results against our standard AOV of around $70 for first time AOV.

At nearly 3x the AOV these outliers tend to provide false positives to the success of an ad.

This is especially true of most ecommerce businesses which I'd venture spend less than $50k per month.

Taking a step back I think it's important that we understand the answers to a few questions:

1. Is Facebook still driving quality traffic? Yes.

2. Is it as much as they used to? Maybe not.

The point of this post is to help you figure out where the quality traffic is coming from relative to your ads and messaging.

That's how you give yourself the best shot of success today.

So how do you know what makes up quality traffic?

This has long since been my question in trying to build out systems to not only answer this question but have a comprehensive understanding of what I can and can't control relevant to driving that same kind of traffic.

The proper elements of a Facebook ad are, claim, product, offer, call to action.

This is the perfect anatomy of an ad.

In fact in surveying people that came from our ads the breakdown was:

QUESTION: What about our ads made you visit?

So this becomes the starting point of understanding this audience that we're driving, all these answers were collected in exchange for a standard popup discount on our landing pages.

So Facebook is sending us a good amount of traffic based on people that are actually looking for products like ours. Their algorithms are working.

They just likely aren't as strong as they used to be in their ability to target people as they once were.

So instead what we need to do is to understand what the correlation is between answers provided about their customer journey and their conversion that can influence us to create ad angles around which audiences are more likely than others to convert.

The above popup asks a series of questions between 4-5 beyond an email address to better understand their customer journey.

1. Which products they are interested in

2. What matters most to them in those products

3. When are you looking to upgrade

Of all the answers collected number 3 is the most important, questions that relate to timing and or knowledge provide the biggest variation in conversion percentages.

The top intent of looking to buy today converts at more than double of the next two variables.

We can repeat this logic across all of the other questions or in combination with multiple answers.

Essentially though, given the skewed nature of the data, we would want to put more money behind ads that are driving traffic relevant to TODAY as the answer than ads where the makeup is more weighted towards other answers.

So these data points, need to be statistically significant to make these judgement calls.

Turns out they are:

First Time Order Breakdowns:

So that means in our given setup that 60.52% of all first time orders have data similar to the charts above attached to them.

So when we look to build out strategies related to this data, we know that we're building toward just under 2/3rds of all first time orders. It's a 50/50 split between people that are ready to buy today and those that will purchase over a period of time.

We're not looking for absolutes, but instead we're looking for patterns among the data where we can better identify the top audiences.

For example if there's a specific product that when segmented skews these numbers even further, then that's a product we want to advertise to drive traffic onto the website.

The above is a data-driven approach to scaling ads in a way that is united with business goals.

When looking at these combinations it allows us to rate the quality of audience coming from a campaign, ad set, or individual ad.

This is regardless of campaign setups, it's not about looking at CTR, CPM, or anything else.

It's simply a matter of driving traffic to collecting data in exchange for an offer as a sign of intent towards making a purchase.

Subscription to conversion rates are 30% overall, 20% from those popups related to ads.

This framework isn't based on account setups, isn't based on any hacks, but actually based on collecting and leveraging data in a strategic manner to design and develop creative that speaks to the quality of audience that is most likely to bring you revenue and orders.

With much less testing and need to create hundreds of pieces of creative.

You don't need a fully polished approach to do this right either, you just need to follow a framework that revolves around three major prongs:

Create an Ad, push it to Facebook, optimize your popup offer to mimic the offer in the ad to require a sign up to unlock it and collect data relevant to the customer journey.

Evaluate the data and then make decisions that are statistically likely to drive results.

I'll tell you why:

Canva - Cheap, easy to use, all ads should have a combination of words and a product, think memes or things that make people question what they are currently using and convince them to investigate your solution.

Facebook for Ad Distribution - it still works, you just need to get the right ad with the right message to the right people, if you create an ad to the proper audience you should find success.

Popup Software (we use Formtoro) - this allows you to build out popups and has built in reporting and a Facebook integration to understand the context behind qualitative and quantitative data that you collect.

Here's how it works:

1. Find a few angles that will not only get your product noticed but angles that will get people talking about your product, comments and likes are the fuel that ads need to reach a broader audience

2. Launch some ads, direct the traffic to a series of landing pages or product pages

3. Set up offer to show from the utm traffic source e.g. utm_source=facebook_ads, utm_medium=paid

4. That way you're only showing the offer from the ads to people that are visiting your website from those ads and can be identified as such

Cool, you've got all the elements you need to add more data to your audience and rank their quality to build out ad angles that are statistically proven to drive results.

The vast majority of brands/agencies do not go this deep into being data-driven, instead paying attention to creatives, hooks, ROAS, and other elements, which are all short term and for a lot of smaller brands with low budgets things that can be wildly unpredictable.

It's true that with the tools today that a lot of media buying has become automated, but just because Facebook's algorithm is automating it doesn't mean it's doing it correctly.

It lacks intent data from a combination of data collected based on where someone is in the customer journey.

Facebook ads used to be easy, the ability to post ads has become easy, but the insights needed to understand not only what's working but what you should be creating before you need to test them is what's missing.

Thanks for reading, will this solve all of Facebook's problems? No.

But is it immune to any future changes or tracking issues that might happen in the future? Yes.

Because it's marketing and customer research insights being applied based on quantitative and qualitative data.

I find myself at a crossroads of wanting to open up this ability to others en mass or just leveraging it for my own personal gain.

The truth is running a brand at 35% net margins is possible. We’re doing it.

There’s a large part of me that thinks that data collection and analytics is just too much for most people at this point.

Because I know this, I have a few options:

Upskill people into taking this sort of approach

Open an agency to run ads using this formula

Partner with an agency to teach them how to do this and take a consulting fee

Take equity in a brands and use our proprietary software to grow them profitably partnering directly with manufacturers

You tell me, what would you do if you were in my shoes?

Separately, it’s become apparent from the past few weeks that the immediate education gap stems from people needing education in these things.

So I’m going to double down on building out some more courses around data collection and analytics to help bridge this gap and lean more into the digital products space.

I’m actively talking to people about taking over management of that side of the business, if this is of interest to you, drop me an email.

I’m currently reviewing proposals.

The Takeaway

Have a great week!

-Jon

Catch up on past posts: https://ecomwithjon.beehiiv.com/

You can learn from me: jonivanco.com