This is great Taylor. What I got from both the analysis is that we should have at least 30 days of data and from that 30 days we should be taking the first 22 days. As the attribution is taking almost 8 days (or even more) to get the right sales numbers. Or analyse the first 30 days of data on the 38th day of the campaign to have as accurate data as possible. Of course, there will be a lot of variations based on product to product. Thanks for sharing this. . .
Here is my understanding of video. 1. last 7 days shows inaccuracy in data, high acos and low sales. 2. Last 14 days data shows accuracy in acos but 2x fold sales. 3. Last 21 days shows accuracy date Acos and accuracy in sales. 4. Last 30 days shows better accuracy in acos and sales.
You mention this data is off of sponsored brand ads and keyword targeted ads. Do you have any recommendations for ASIN targeted ads in terms of time frame for bid adjustments? Is it the same? Thanks :)
Hey! Destaney directed me to this video. I am a first time seller and have sold out in about a month based off limited PPC (although I plan on launching more and more campaigns as I scale properly). After watching this video and speaking with Destaney I have taken away the following: When I go in to optimize my PPC I should base it on 30 days worth of data. So on October 1st, I'll look at my ppc numbers based on Sep 1-28th and make bid changes... Then 3 days later or so I'll optimize again based off the numbers from Sep 4th-Oct 1st etc etc? So basically a month's worth of numbers minus the last two days because of attribution/refunds and reoptimize every 3 to 4 days? Does this sound like a decent strategy for a small private label seller or no? Thanks!
If I am not wrong, and look at your sales attribution window test, it appears to me that at the end of 14 days, almost all of accounts (38 of them) have 0% disparity comparing to the data you pulled on the first day. It leaves me to believe that the data would be more accurate if we pull it every fortnightly as it may also indicate that Amazon would settle payment every fortnightly to the sellers. Same phenomenon happens to your big five accounts, as ACOS is very similar when reading from fortnightly basis. But I don't understand is that there are vast differences between 21 days sales and 14 days sales, and it doesn't make any sense even though we can factor return and later purchase.
The 14 day number being significantly higher in spend and sales doesn't make sense. In the first example, going from ~$20k spend to ~$97k spend down to ~$44k spend and finally up to ~$61k spend 30 days makes me wonder if somehow Amazon's reporting is presenting incorrect information. Have you tried connecting back to the Amazon Advertising invoices? How would Amazon account for spend changes like that?
Hey Devin!! That's a good variable to cross reference with I hadn't thought about it like that! Our internal rep had told us this is normal & that there's 3 major variables we need to consider 1. Sales attribution 2. Refunds 3. Holidays ( since the data was basically covering the December timeframe ) I'll have to speak with him about how Amazon would account for spend changes, I imagine it's something similar to click fraud where they give a refund or just deduct the amount from the next invoice, but we'll see what our rep says on our next call! Thanks for the Question :)
Wow! I never considered the attribution window could be so big! Thank you sm for sharing this
Thanks Taylor. It's hard to resist that urge to optimize bids too soon, but this vid helps put it in perspective and will help us be more patient.
This is great Taylor. What I got from both the analysis is that we should have at least 30 days of data and from that 30 days we should be taking the first 22 days. As the attribution is taking almost 8 days (or even more) to get the right sales numbers. Or analyse the first 30 days of data on the 38th day of the campaign to have as accurate data as possible. Of course, there will be a lot of variations based on product to product. Thanks for sharing this. . .
Here is my understanding of video.
1. last 7 days shows inaccuracy in data, high acos and low sales.
2. Last 14 days data shows accuracy in acos but 2x fold sales.
3. Last 21 days shows accuracy date Acos and accuracy in sales.
4. Last 30 days shows better accuracy in acos and sales.
How reliable is the data in Q4? can we see some trends in other Qs?
You mention this data is off of sponsored brand ads and keyword targeted ads. Do you have any recommendations for ASIN targeted ads in terms of time frame for bid adjustments? Is it the same? Thanks :)
Hey! Destaney directed me to this video. I am a first time seller and have sold out in about a month based off limited PPC (although I plan on launching more and more campaigns as I scale properly). After watching this video and speaking with Destaney I have taken away the following: When I go in to optimize my PPC I should base it on 30 days worth of data. So on October 1st, I'll look at my ppc numbers based on Sep 1-28th and make bid changes... Then 3 days later or so I'll optimize again based off the numbers from Sep 4th-Oct 1st etc etc? So basically a month's worth of numbers minus the last two days because of attribution/refunds and reoptimize every 3 to 4 days? Does this sound like a decent strategy for a small private label seller or no? Thanks!
Very Helpful. Thank you for this information and confirming the timing. Much Appreciated!
For sure!! glad you found value in it!
Extremely helpful thanks for the info !
Glad you liked it! that's great to hear :)
If I am not wrong, and look at your sales attribution window test, it appears to me that at the end of 14 days, almost all of accounts (38 of them) have 0% disparity comparing to the data you pulled on the first day. It leaves me to believe that the data would be more accurate if we pull it every fortnightly as it may also indicate that Amazon would settle payment every fortnightly to the sellers. Same phenomenon happens to your big five accounts, as ACOS is very similar when reading from fortnightly basis. But I don't understand is that there are vast differences between 21 days sales and 14 days sales, and it doesn't make any sense even though we can factor return and later purchase.
The 14 day number being significantly higher in spend and sales doesn't make sense. In the first example, going from ~$20k spend to ~$97k spend down to ~$44k spend and finally up to ~$61k spend 30 days makes me wonder if somehow Amazon's reporting is presenting incorrect information. Have you tried connecting back to the Amazon Advertising invoices? How would Amazon account for spend changes like that?
Hey Devin!!
That's a good variable to cross reference with I hadn't thought about it like that!
Our internal rep had told us this is normal & that there's 3 major variables we need to consider
1. Sales attribution
2. Refunds
3. Holidays ( since the data was basically covering the December timeframe )
I'll have to speak with him about how Amazon would account for spend changes, I imagine it's something similar to click fraud where they give a refund or just deduct the amount from the next invoice, but we'll see what our rep says on our next call!
Thanks for the Question :)
Very helpful, thank you! Will try to be more careful with the main keywords. And make conclusions based on 30 days data. Thanks a lot for your time!
Exactly! glad you liked it