Google ads tracking issue how to analyze

[frankthetank]

Hi there,

I was given a task to analyze the google ad campaigns in excel and one of the notes attached to this task said that:

"For every booking received from AdWords, the company receives an additional 0.75 bookings. This is due to tracking issues related to cookies and multi-touch attribution."

My question is, how would this impact the data? The files that I was given are: Leads, Bookings, and Campaign reports (daily and summarized).

As I was given the google ad data (as mentioned above), I am not too sure how to approach the "factoring in" of the above-described tracking issue. Does it have to be considered? If so then how? The goal of the campaign is to create a conversion funnel to help decide on the fund allocation to the most successful campaigns.

Many thanks!

[RhinoFish]

It doesn't impact the data, it impacts the conclusions drawn from the data.

If you track things as best you can, but know from turn-off-the-ads tests that sales shrink by 75% MORE than previously reported PPC revenue when the PPC is off, and if you know this is relatively constant (usually not) across branded and unbranded AND Search and Shopping, then when you set ROAS Goals, you can assume that the revenue that you track follows this math: Tracked PPC Revenue x 1.75 = Actual PPC Revenue

This gap between Tracked and Actual ALWAYS exists (no tracking system is perfect), but if you know the approximate size of the Tracked / Actual gap, you can allow for it when you set your goals.

If you start prepending "Tracked" or "Actual" to numbers, it'll help clarify.

Tracked PPC ROAS = 5.0x
Tracked PPC Spend = Actual Spend = $10,000
Tracked PPC Revenue = $50,000 [Spend x ROAS]
Actual PPC Revenue = $87,500 [Actual = Tracked x 1.75]
Actual PPC ROAS = 8.75x [Actual = Tracked x 1.75]

I used the word "Revenue", you can sub in whatever outcome you're producing... Revenue, Sales, Bookings, etc.

[frankthetank]

Hello! Thanks for the response.

Sorry, just to clarify few things - Are you saying that the number of conversions (as per the campaign reporting) does not come through as inflated? i.e. for one ACTUAL booking, the company receives an additional 0.75; doesn't it mean that the 0.75 is flawed due to tracking issues and therefore, the actual Google Ad campaign output should be adjusted as per below (sorry about the format)

Budget:710
Budget type:Daily
Status:Eligible
Impressions:405279
Clicks:54860
CTR:13,54%
Avg. CPC:2,73
Conversions:557,33 (isnt this 1,75 of the actual?)
Cost:149757,76
Adjusted Sum of Conversions:318 (should not this be adjusted - 100% out of the 175% above?)
Adjusted Cost per conversion:471


Another question that I have is that I received a separate file with sales/bookings and (unlike when it comes to conversions above) the number does not have decimals (hence why I assume it reflects reality as there are also details of bookings included) - a) what is the relationship between the sales/bookings and the above conversions as both come from the same campaign (which I have no further details about) ? How come the conversion number is much higher than sales (roughly 300 conversions on 100 actual bookings) and should only the revenue from the actual bookings be used to calculate revenue? Or should conversions be used for such calculation as well?

Also, once again - can you please confirm how to factor in the ""For every booking received from AdWords, the company receives an additional 0.75 bookings. This is due to tracking issues related to cookies and multi-touch attribution." ?

I see that in your example you actually inflated the ROAS by Adding the extra 0.75 while what I understand is that the 0.75 comes through as an error.

Thank you VERY MUCH for your response.

[frankthetank]

Last thing - would the CPA be calculated using the conversions or the actual (physical) bookings received by the company (where the Analytics Campaign (First Session) = the evaluated campaign) ?

Thanks

[frankthetank]

Hi again and my apologies for bombarding you! I gave it a final thought and I think I understand what you're saying. While the conversions (and any other data received from Google Ads should stay untouched (unlike what I said above), the sales related to the respective campaign should be factored by 1.75 (as the tracking issues result in 0.75 booking missing/i.e. not being correctly associated with the given campaign as it should). So in result the booking received and associated to campaign A should be x1.75 and so should be the associated direct cost per booking - unlike the PPC costs as these are reported correctly. Furthermore - should I also assume that the received leads are also subjected to the same tracking issue (i.e. need to be x 1.75? or should this not be assumed unless specified?). Last thing - should the CPA be calculated using conversions or the actual bookings/sales (in this case x 1.75)? Kindly confirm.

Thank you so much - this is indeed very helpful way to comprehend the way Google Ads (and its respectively) works.

Kind regards,

Frank

[RhinoFish]

Yep, you figured it out.

When they said this:
"This is due to tracking issues related to cookies and multi-touch attribution."

They were saying things like...
1) Tuesday the shopper clicks on an ad, an ad blocker or browser security blocks the cookie, they return tomorrow, type in the domain name, place an order = the cookie isn't there, the order does not track to PPC.
2) If importing Google Analytics eComm conversions into Google Ads as the source of tracked Google Ads Conversions, in Google Analytics Last Touch gets 100% credit for the sale, so they click on a PPC ad, last minute they google "merchant name coupon" to grab a discount code, the coupon site forces the user to click on a reveal codes box, this causes the merchants store to load in a new tab... so Google Analytics sees TWO touches for this sale, the original PPC click, THEN an affiliate coupon site click, and Analytics gives 100% of the sale credit to the Affiliate coupon site touch, so there will not a a Google Ads driven conversion imported from Google Analytics into Google Ads, because Google Analytics uses Mutli-Touch attribution AND they default it to Last In, so according to Analytics this sale is not a PPC conversion.

Because of ad blockers, things that break cookies, security moves made by modern browsers, and other factors, Tracked PPC Sales is ALWAYS less than Actual PPC Sales.

The only way to absolutely prove this is to turn off the PPC and see how far sales drop.
If PPC tracks as having driven 33% of sales, but when you turn off all PPC, sales drop by 47%, you will understand that not all Actual PPC Sales were actually Tracking.

:-)

[RhinoFish]

In Google Analytics, you'll see, for example, a category of Source / Medium called "Direct / (not set)".
These are the cases of purchases where Analytics did not track anything as having driven a sale.
You know that people don't just wake up and type CandyMart.com (I made up that domain name) into their browser... something drive them to go to the store...
The Direct / Not Set comes from bookmarks, direct types ins, blocked cookies and blocked header hand offs.
AND, Direct / Not Set is the JUST the bucket where the actual source was blocked AND there were no other identified sources.

Imaging a string of touches that looks like this...
[day0, PPC click], [day0, 3 hours later, Shopper typed in your domain name], [two days later, day2, shopper types "Candy Mart" into google.com and clicks on the organic listing there], [day4, Shopper types in domain name, signs up for newsletter cuz you flag it as a wellspring of discounts], [day7, you email them a 5% off coupon and they click on the email link and buy]

Sorting out credit is hard enough... now imagine the shopper uses an ad blocker that nuked that day0 PPC cookie... that PPC click is invisible to all attribution systems.

There will also be cases where PPC is over attributed, but in general, your unbranded, early, upper funnel PPC touches, are vastly under counted.

And, lastly, you very likely would have NEVER captured their email if that PPC click didn't happen. Yet your email platform will mark it as having driven the sale.

Hahaha! Every channel as a reporting silo will claim credit. So if you add up the sales in every channel, then compare the total to the sales figures in Analytics (that tracks all channels), it is likely that siloA + siloB + siloC + siloD >> Analytics

What a mess. Good news though, everyone you compete with faces this same challenge! If you do a better job of unwinding it, you've found a great competitive advantage over those you seek to pummel and bruise.
:-)

[frankthetank]

Fantastic, thank you SO much for the explanation, really helpful! What about the CPA and the leads though, should I, in this case, use conversions or the actual sales in order to calculate the CPA? Also, should I also assume that the received leads are subjected to the same tracking issue (i.e. need to be x 1.75?. Or should this not be assumed unless specified?). Thank you so much once again!

[RhinoFish]

"should I also assume that the received leads are subjected to the same tracking issue"
Anything tracked by page-mounted / app-mounted scripts suffers from this difference between Tracked and Actual, so yes, Lead conversions and Purchase conversions are affected.
On the input side of things, like Imps, Clicks, CPC, and Spend, Google's ad servers track that as it happens, these numbers are then Actual = Tracked = Reported.

"should I, in this case, use conversions or the actual sales in order to calculate the CPA"
Since your client already understands that there is a difference between Actual and Tracked, you should probably discuss both with them.
Like I said, get in the habit of labeling the outcomes (Leads, Sales, CPA, Revenue, etc) as either "Tracked" or "Actual".

And consider yourself lucky to have a client who knows there is a difference!

[frankthetank]

Thank you RhinoFish - great to learn a bit more. Since this was a mere exercise, I have just included a note that I skipped the adjusting (i.e. x1.75) of the Leads and Conversions (despite the tracking issue likely having a similar effect on those) and only adjusted the Sales (as per the instructions that I mentioned above) and the related direct costs. Hope that is not a huge deal - even though it will push the CPL and CPA (should the conversions be used in place of sales) up. Thanks a lot for all your help - and let me know what you think of the approach that I have taken.

[RhinoFish]

"and let me know what you think of the approach that I have taken"
The client told you to consider the gap, and you did - as long as the client sees that you understand their input and that you listened and planned using the info they provided, what I think is out the window, serve the client's needs and all will work out best.
:-)

Since the client knows that the gap is important, I still think it would be a healthy thing to label things as "Actual" versus "Tracked" / "Reported" when giving them reports.

The Tracking vs Actual gap is a big deal, bigger still is the attribution model being used to make decisions.

A great biz friend of mine, after listening to me discuss Tracking vs Actual AND Attribution Modeling for many months, coined the phrase "Great Tracking and Attribution Modeling is a Competitive Advantage". Sooooooooo true!

Hope you and your client leverage your tracking know how into body blows and upper cuts landed on those you compete with.
:-)