A [very small] Case Study on Zazzle MP Search Results Ranking

[ColsCreations]

This page
What Royalty Rate Do You Recommend I Use?
suggests a royalty of 5-10% which is pathetically low, but if you read the whole entry and between the lines, it’s clear that

1. it’s not mandatory, you can set your royalty to whatever you want
2. lower royalties = lower overall prices = Zazzle looks more affordable to shoppers
3. using that low 5-10% benefits the customer not you

There’s a common theory that higher royalties might keep one down in search results. So i just wanted to do a random search to see what royalty percents were in the top results. That by itself proved interesting and then it spiraled into a deeper dive looking at the other factors that most likely weigh in to where something “places” in results. 

I searched (on the US domain) for three keywords + product type. I.E. [keyword1 keyword2 keyword3 product_type] NO QUOTES, and then used the “Refine by Category” filter to narrow results down to just that actual product type. That gave me just over 200 results, a number easy to manage for comparison purposes. 
I then noted the details for the top 20 results. In the four days I’ve been compiling the stats on the results, the search results fortunately remained exactly the same. And then I was given the idea to do the same search on other domains. So I did UK and Canada and noted where items placed there compared to searching on the US domain. 

Of the top 20 results, I was only able to obtain all the stats for 7 of them. That’s 35%. Not enough to draw any “scientific” conclusions, but still super interesting and able to provide some insight. 

The first thing to jump out is that in most cases there is an astounding difference in placement depending on domain! 

Designers of all Pro levels were included in the top 20. There were two with a generic Pro badge which means they could be anywhere from Basic to Diamond. None of the results were from Designers without at least the basic Pro badge. Does that mean anything? I honestly have no idea. 

Royalties ranged from 10 to 25%.
Views ranged from 11 to 347.
Creation dates ranged from 2014 to 2024.

Looking at the meta data:

* All of the results have all three of my search keywords somewhere in the meta data.
* One of the top 20 (#16) has all three keywords in the title, in the same order I typed them in search
* Five of the top 20 have two of the three keywords I searched for in their title, in the same order I typed them in
* Ten of the top 20 have none of my keywords in their title, they only appeared in the tags

This list is missing 65% of the data but is still interesting and my own personal conclusions, right or wrong, are, in no particular order::

* number of views alone doesn’t seem to be a heavily weighted factor
* even number sold won’t by itself put you at top of results
* likewise, Pro level or zRank alone won’t put you at the top
* having a personal cover image is still relevant/important
* royalty % by itself is not keeping items down in search
* last sold date might be an extra-weighted factor with more recently sold items placing better
* where your strongest keywords appear and in what order compared to what a shopper types in is important

So that’s it. My conclusions here are just my own opinions based on the limited data of this search, you all may draw your own different conclusions. 

[MOM]

@ColsCreations  Very insightful - thank you so much!!

[Windy]

Fascinating! I think you are saying that only two of these products have personal cover photos. That surprises me.

I am wondering how zRank and number sold are known… would I be able to find that information also? If I knew where to look? 

[ColsCreations]

I reached out to some of the Designers who had products in these top 20 and they were willing to share to help create a bigger picture we all might be able to learn something from. So no worries, no way I know of for anyone to see those stats if not voluntarily shared.

[ColsCreations]

Yes, of the top 20 only #1 and #16 had personal cover photos. Of the other approx 200 results, only 5 of them had covers. I should take a closer look at the stats on those. I’d guess their keywords weren’t close enough to my query to raise them up even with a cover. But it could be any combination of other factors. Hmmm… 

[ColsCreations]

Also - forgot to note - 

I did the same search numerous times over several days and always got identical results whether I was signed in or using an anonymous browser,  to make sure I wasn’t getting results based on just my own Zazzle browsing history.

[Windy]

Ah! Very good. Extremely helpful information. 

[Mari_Holly]

Wow that is really thorough! Thank you so much for your dedication! What type of product did you search for? I assume it was not a mass product like invitations or cards but something very specific. I say this because I find it very curious that only two had a cover image but do you still think it is relevant or did I misunderstand?

[ColsCreations]

Well, this far in I figured why stop here so went to look at the others with cover images (6) that weren’t in the Top 20.  

I started this search on Aug 9. On the US domain, there are 239 results, 204 after narrowing to just specific product type using the category filter. Of those 204 results, a total of only 8 have a personal cover image. These totals are pretty much the same on the UK and CA domains. 

Results for this search have been consistently the same all week each time I did the search, signed in or anonymously. Until tonight, 7 days later, there is one new appearance in the Top 20 on the US domain. Every other result is the same, in the same order, except this new one now comes in at #18, so the former 18, 19, & 20 are now pushed down to being 19, 20, & 21.   (#1 and #16 in red are the only two with covers.)

The new #18 (highlighted yellow here) does not have a view count on the product page. I don’t know what that means. Is there a minimum number of views necessary for the counter to appear?
I have no idea where #18 placed prior to tonight so no idea why it’s now jumped into the Top 20.

Had to note this small change to the list before updating further …

[ColsCreations]

Of the total 204 results on US domain, here are the other 6  that have cover images:

 Three of them (30, 138, 142) have drastically different placement on the two international domains than they do on the US domain. The other three (132, 145, 191) are pretty close to where they place on the US domain. In all six cases though the placement on the two international domains is almost identical.  It’s becoming clear that the search algorithm is weighted differently depending on which domain a shopper is searching on. No idea what makes the difference.  ???

One thing I noticed for the first time this week is that total views reported is the same no matter which domain you are viewing a product on. So it’s not differentiating between say, you had 20 views on the US domain, 5 on the UK domain, and 0 on the CA domain… It’s just showing 25 total views no matter which domain you are on. It would be really cool if when viewing from a different domain it only showed us the views from that domain. That would help us see where our items are generating the most interest from and we could build on that when designing, marketing & promoting. 

Anyways, tomorrow I’m going to look closer at the keywords used in these Top 20 and the other six that have covers but still didn’t make it into the Top 20. Could all be for naught, maybe Z search just throws in random items for no identifiable reason. I don’t think we’ll ever figure out the secret sauce but it’s been an interesting look so far. 

[Anne]

Thanks so much for sharing all your research, @ColsCreations  
I have a question about this statement:
* having a personal cover image is still relevant/important
If only 2 out of 20 (from your first table) have a cover image (and subsequent research), I would say it is not so relevant for ranking?

What I find disheartening is that all of these product were created almost or more than a year ago. I don’t know what product and tags you used, but maybe that means if you would search for something more trendy, you would get more recent results? Or should we conclude that it does not pay to create something trendy as only the same old is shown first? I am also assuming that this is new, as I noticed it this year that nothing created in the past year is selling (and that is in contrast to previous years).
https://community.zazzle.com/t5/ask-a-question/search-engine-stuck-in-time/td-p/214053
It seems to me to be a counterproductive strategy to only place old designs in the top shown?

[Sea-Change]

This is great info, @ColsCreations ! Thank you for going the extra mile to fill out additional details. It provides helpful insight on royalties and much more.

Adding on to others’ obervations on the dates, I noticed that 9 of the top 20 results are more than 5 years old, in addition to the lack of 2025 designs. I realize that indexing can take a while, but I would think at least 1 design would have surfaced within 8 months although perhaps this varies by search.

I have noticed the same staleness in some of my casual searches; for example, I’ve seen little change in the past 3 years on the first page of business card results and in many searches. There are so many great business card designs that don’t get much exposure.

[Mari_Holly]

Just to comment: From my few recent sales it turns out that there are new designs. I couldn’t say right now how many, but several. This drives me crazy because I don’t know how to interpret absolutely anything that’s been happening lately with sales.

[ColsCreations]

I have a question about this statement:
* having a personal cover image is still relevant/important
If only 2 out of 20 (from your first table) have a cover image (and subsequent research), I would say it is not so relevant for ranking?

#16 is my item. So I knew it had no sales, few views, and a high royalty. Which is why I used it as my test-search item. I was really surprised it came up in the top 20 given those stats plus being only Bronze and zRank of only 6. So my first thought was, has to be that it has a cover - why else would it not be at the bottom of the pile?

But as I started looking/pondering more, out of the 204 results, there were 6 others with covers that weren’t in the top 20 and only one of those even made the first page of results. The other five were way down, on the 3rd page.  So right, if covers by themselves give you a boost, those six should have come in higher. 

Then a quick look at the keywords, mine was the only one of the 20 to have all three of the keywords I used in its Title, in the same order I typed them in to search. So I thought, OK, that’s why it placed where it did, not just because it had a cover. But 10 of the Top 20 didn’t have any of the keywords in their Title, they were just in the tags. So clearly a lot of different factors in the algorithm and I’m going to look more closely at the keywords of the rest of the results, especially those other 6 with covers, to see if there’s any kind of pattern there. So my mindset today - which might change tomorrow - is that covers are still relevant/helpful when combined with other factors but they’re not going to rocket you to the top by themselves. 

[Anne]

Thank you, @ColsCreations  
“why else would it not be at the bottom of the pile?” 
When I read your description it reminds me of my own searches for products which always show up one of mine on the first page. Even after clearing cache, having a privacy cleaning program running and what you have. I suspect somehow Zazzle’s pixel technology or whatever is sneaking through and your search results reflect something of your own Zazzle account.
Also Zazzle has “favorite designers” whose products will always rank very high, regardless of other factors.
Any way, your research is interesting, so thank you for that!

[ColsCreations]

When I read your description it reminds me of my own searches for products which always show up one of mine on the first page. Even after clearing cache, having a privacy cleaning program running and what you have. I suspect somehow Zazzle’s pixel technology or whatever is sneaking through and your search results reflect something of your own Zazzle account.

Agree and that’s why I did the search in different browser w/ everything cleared. But still suspicion that it may somehow be reflecting my own Zazzle history. But another Designer doing same search on their end got same results. And because they are not in the US they thought to do the same search on their local domain and when I then searched on other domain, I got same results as they did. So that’s what clued us in to placement varying so wildly depending on domain. 

[ColsCreations]

So clearly a lot of different factors in the algorithm and I’m going to look more closely at the keywords of the rest of the results, especially those other 6 with covers, to see if there’s any kind of pattern there.

I finally did this a couple of nights ago. 

The new chart below is the original list (but including the “new” #18 that appeared when I last looked at all this on the 16th.) 

* The yellow #16 is my product and what made me start going hmmmm… 

* The 7 with gray backgrounds are the other version of this product that is higher priced to begin with. Thus why the price is higher than those with lower royalties. But they showed up in the Top 20 regardless. 

* The 8 red numbers are the only ones out of all 204 results that had cover images.

* The 2 question marks (?) in the Views column are products that oddly don’t have the views shown on the product page, not even looking again today. 

* For this search I used three keywords (K1 K2 K3) + the product type which happened to be two words. All 204 results had that product type auto-appended to the end of the Title. If the product type was also included in the tags, I noted that, with Partial meaning the first of the two-word product type was used somewhere in the tags. 

* the …  's represents where one or more other words (that weren’t the keywords or product type) appeared. Hopefully that makes sense and you can see from the chart where the keywords appeared and in what order and if they were immediately adjacent to each other or had other words inbetween.

Still not enough real data to draw any “scientific” conclusion from, but it’s something. 

I didn’t look at the meta on the other 177 results, but for these 27, they all had all 3 keywords somewhere in the tags or titles. And all 27 were assigned to the same category, except for #13 which was a loner being in the Wedding cat. 
#20, #21, and #30 had what I thought were kind of spammy tags for different reasons. 

My #16 is the only one to have all 3 keywords in the Title and just as I typed them. #12 is the only other to have all 3 in the Title, but they aren’t in the same order I typed them into search, and other words come between them. It also has no cover but my guess is that it’s got some sales on it so that’s why, like the others, it’s above mine with no sales but exact keyword match in title. 

That’s all I’ve got today. Make of it what you will.

[GagCartoonist]

Based on this dataset, likely ranking factors (weighted roughly in importance):

1. Sales volume and conversion rate (most important)

2. Price competitiveness (mid-range sweet spot)

3. zRank (as a threshold / credibility measure)

4. Listing recency (boost for fresh designs)

5. Engagement (views & likes, but only if they lead to conversions)

6. Regional performance (localized algorithm differences)

7. Royalty (indirect effect via price, not a direct ranking penalty)

re sales and social proof

• Best sellers:

• No sales products:

Actual sales volume (and maybe recency of sales) appears to be a significant factor. Even a small number of sales can push a product much higher than one with none. 

[ColsCreations]

I didn’t mention the actual product type or specific keywords I used as I didn’t want people to be able to figure out who the designers who shared their stats anonymously are. I’d consider it a common item though but I specifically chose something that isn’t in the invite/stationery category or clothes as they have too many variables you can publish as which would have made it very difficult to compare results.  The keywords I used were fairly common I think, not particularly niche, and I used three to keep from having thousands of results. 

Now that it’s been two and half months, I’m going to look and see where the top 20 place today. 

[ColsCreations]

Did a new search today to see how results may have shifted since first digging into this in Aug.
Still getting 204 results as before after using category filter to narrow down to product type. 

The #1 result is still the #1 result. Everything else shifted some.
5 products (the blue numbers) moved up a bit.
The other 15 (red numbers) moved down.
Only three (the former 16, 18 and 20) fell out of the top 20 altogether.
(I’m showing the top 21 results here because with my last update in Aug, a new item had moved into the original top 20 list, pushing what had been 18, 19 & 20 down to 19, 20 & 21.)

The newcomers to the list are the items currently at #9, #14 and #21, shown at the bottom here. I have no idea where these products placed before today. But #14 - a very old product with no view counter showing on the page and yes, a 50% royalty rate. Trying to see if a higher than average royalty would keep you down in search is why I started this project. There are clearly many variables that together determine where you land in a search, but I’m really feeling that a high royalty by itself is not going to keep you down.  

I think the biggest change here in the orig top 20 is #16 dropping to #35. And because that’s my product I think I have a plausible explanation for that. My store had been at zRank 6 but then after a big self-referral sale was cancelled on Aug 19, my zRank immediately was dropped to 5. (Side rant here: I was at 6 when the sale came thru on 7/3. That sale & referral did NOT increase my zRank, it stayed at same 6. But then the cancellation dropped it to 5. And that’s just dirty. If the sale had increased my rank to 7, it would be fair/understandable to drop it back to 6 with the cancellation. But they dropped my rank for the cancellation even though it it had never been raised with the sale. ) I’ve always believed that zRank by itself is not all that important in search results, keywords are. But here, the drop in zRank is the only explanation I can think of for the slide from #16 to $35. 

The other notable change was #12 moving up to #3. Few new views on it, so my guess is the product sold since my searching in Aug and thus it’s now placing higher. 

The other thing we were pondering here was how important covers are. Of the orig top 20 list, only #1 and #16 (now #35) had/have covers. Out of 204 total results, there were only six others that had covers. All six of those have dropped some in placement with the former #30 seeing a remarkable drop down to #107. There are three products that now have covers where they didn’t before but like the others, none of them made even the first page of results, not even the one from a Diamond with 436 views. 

It pains me to say it, but while I think it’s important to add covers, I don’t think a cover by itself is going to rocket you to the top of results. 

Other notes:
1) I did this search both as me signed-in and in an alternative browser, everything cleared, not signed-in, and got same exact results. So the results I’m seeing are not being skewed by my own browsing history on Z. 
2) Not a single product on this list has changed in price, meaning the drops in “rank” can’t be attributed to the designer having changed their royalty on them. 
3) the couple of items with a gray background are ones that were published as the more expensive product option, that’s why even with a lower royalty rate they have a higher price
4) #19 (now #20) is showing an alternate color version in search results than it was before. 
5) the number sold / last sold dates have not been updated, did not want to bother the other designers who shared their orig info to see if their item has had any more sales since.