Your Facebook ROAS is Undercounted. Here's what to do about it.

Your Facebook ROAS is Undercounted. Here's the solution.​

Too good to be true, attribution and blended results

Too good to be true, attribution and blended results

Seeing a lower ROAS than expected on Facebook?

Not all of your purchases are being attributed by Facebook. The very least your Facebook ROAS is undercounted.

This is because of the Herrmann Hypothesis. One Marketer named David Herrmann noticed a trend between Apple’s latest Software Patch with the iPhone 11. In summary, Apple is blocking cookies by default after 24 hours. 

This means the purchase pixel isn’t firing and Facebook isn’t attributing past a 1 day click 1 day view.

One way to get around this is offline conversions. 

I’ve used offline conversion data in the past but that was only for brick-and-mortar stores.

Never even thought this could be used for online sales until Depesh Mandalia mentioned it.

He wrote a post in September 2019 explaining why Facebook ads were not attributing. He recommended uploading offline data as one of the solutions.

At first sight I was stunned.

In one account I thought I was tanking hard. But it went from 1.72 to 3.0 ROAS by loading it!

1.74x multiplier in increased ROAS.

The account in question
Another account that had similar increase in ROAS. From 2.25 to 3.41.

I was sharing it around the Facebook Paid Media community saying “I highly recommend you upload offline conversion purchase data into your events manager ASAP.”

But it turned out too good to be true.

It was first brought to my attention by a fellow media buyer, Casey Carroll, that offline event sales may also include purchase pixel sales. So there may be duplicate sales and the ROAS is inflated.

When you select the purchase event code for the offline event set, it will also add the same duplicated sales numbers into the overall purchase event too.

A few other keen marketers provided this insight too.

What this means is Facebook attribution is as tricky as ever. 

Initially I recommended that during demanding times like Black Friday / Cyber Monday and for big accounts it might make sense to upload offline conversion data hourly if not every 15 minutes.

However, that’s only good for directional data like Jason How and Josh Durham suggest. Making educated guesses on a sliver of the complete data.

Attribution is already complex if you take into account Delayed Attribution Multiplier (DAM). The 1 day click 1 day view will look a lot different from 28 day click 1 day view.

But because we can’t wait 28 days, we have to project it based off historical data. Not to mention the possibility of projecting 60 day, 90 day and 120 data. 

But that’s only a piece of the puzzle. Not to mention omni-channel attribution and different attribution models (last click, positional, time-decay). Who gets the credit!?

For now, the most accurate way to get more accurate Facebook attribution data is de-duplicating overlap between offline purchase data and purchase pixel.

Right now no functionality to do this (yet) but backend must be all there. 

Someone’s going to eventually have to solve this challenge. I’m going to nudge a Facebook Engineer to dig into this. Stay tuned… 

There goes the saying by a old Marketer named John Wanamaker “Half the money I spend on advertising is wasted; the trouble is I don’t know which half.”

It still holds true today.

Practically if we know that our data isn’t fully accurate on Facebook the best approach for performance marketing is taking blended results. Total Sales Revenue divided by all your paid media channels.

That way we’re less focused on our Facebook ROAS and more about our Business ROI.

At the end of the day if purchases across all our ad accounts aren’t being fully attributed to Facebook there still lies a challenge.

Without accurate data, our decision can be affected & conservative. Especially during times to scale and to spend aggressively when we know our numbers to make educated decisions.

That we know what our break even ROAS is based on our Cost of Goods, Sales Discount, and Ad Spend.

Once all those numbers are aligned then we all have the best shot at making data driven decisions.

If not, it’s just a shot in the dark. For now Marketers and Brands have a flashlight by looking at blended results. 

After all, I like my marketing like I like my avocado smoothies: Blended.

Someone will eventually figure out how to turn the street lights on but for now we got to work with what we have.

Get the 5 E-Comm Arbitrage Levers on how to recover lost revenue in just 21 hours.

Whether that’s serving you by helping your CMO, agency, in-house marketing team, or paid media contractor. I help you keep plugged in with e-commerce arbitrage opportunities. You have an entire brand and business to run, and I’m here to help you fill in your R&D gaps.

My aim is having free marketing material that rivals any incubator, strategic partnerships or agencies. There’s information over-load, let me the one to filter it all out and curate best practice for you.

Start by getting the 5-Point Playbook that shows what you can do today to recover lost e-commerce revenue.