The Grand Experiment, How Much Does Free Shipping Affect Traffic?

[rwperu34]

For those of you who just want the answer, it's improves traffic by about 10-15%.

On Sept. 12, shortly after midnight, I reset my inventory by hitting sell similar on every single item so eBay would think they were newly listed items. Then on Sept 19, shortly after midnight I did that again, but this time I added the shipping charge into the price of the item and offered free shipping on 100% of my items. This should do a very good job of isolating the free shipping bump.

I didn't run any auctions. I didn't buy or list any new inventory. The only activity on my account during this time was listing items that sold if I had a backup and a bunch of delisting and lowering of prices on older items. If you're curious why I couldn't put that off, it took the whole two weeks to complete that project. Ideally I wouldn't have done any of that, but I made sure to spread it evenly over the two weeks.

Impressions-From Sept 19-25 my impressions were up 7% for the week. This is on a smaller inventory though. Per item impressions were up closer to 15%. Almost half of the bump (23%) came on the first day (ie the day the items were "new"). After that first day the bump was a more modest 3%.

https://i.imgur.com/ghH4gS0.jpg

Sales-Why not use sales as the barometer since that's what we're really after? A couple of reasons. Reason #1 is sales have a lot more variance in the short run. Over one week traffic is a better predictor of sales than sales itself. Reason #2 is I did not have a method for combined shipping during the free shipping week. The 30% drop during the free shipping week can be 100% attributable to this. Probably 110-115% in reality.

Sell Similar-The sell similar (or new item) bump is HUGE!!! Traffic was up over 40% for the first week when I reset my inventory. At that time I would say about half of my inventory was over a month old and at least 90% was over a week old. The bump was the biggest on the first day (75%) and maintained over 30% for 5 of the next 6 days.

What you can't see in this chart is the Monday to Tuesday bump. Normally Tuesday is less than Monday. By normal I mean 30 out of 30 weeks that weren't affected by vacation or running an experiment Sunday was my biggest traffic day of the week, Monday would be less than Sunday, and Tuesday would be less than Monday. From Monday the 18th to Tuesday the 19th on inventory that was all exactly one week old I saw a 30% bump after hitting sell similar.

Takeaways-The first takeaway is anybody trying to claim you'll double or quadruple your sales by offering free shipping is selling you snake oil. I don't know the exact number, but I can say with certainty that it's not a 100% or more bump. I knew that even before running this experiment based on my sales (I have both free shipping and shipping charge listings).

The biggest takeaway for me is actually the power of the sell similar bump. I knew it was big and thought recycling items to be a worthy use of time, but I'm going to be even more aggressive in the future. Even if I think it's the cheapest, lousiest inventory I've got, I'm not going to let it get past two months. Anything I think that is decent will get recycled within 5-6 weeks at most, and would be considered after one week.

[longhornjunkie]

Great info. Does "sell similar" work in bulk? (I'm not at a well-functioning computer off my company network often enough to just try it.)

Another advantage to free shipping I remembered this week is on ESE items when you file a claim and don't get the shipping refunded. Seems like you should get that back since, you know, that's the service you paid for that failed.

[fungi2510]

Longhorn had the same question I started with. Did you do these individually, or is there a way to do it all at once? I have been trying to do 10 or so twice a day to refresh my inventory, but its a chore.

Second question: did you change any prices when you relisted them? I have been changing my prices to try to increase sales. Obviously it is working a little bit.

[checkoutmydeals]

Was there any change in multiple item sales?

When you sold multiple items to the same person, did you ship in separate envelopes or combine shipping?

If you combined shipping, did you rebate anything to the customer, or throw in any bonus items?

[shrevecity]

What about intangibles that you cannot control. I can look at my impression charts and see impressions fall later in the month. That is one then you cannot control just number of searches for that item being made. You have a card of joe Bob he throws a ball and hits a kid in the stands searches for him will go way up and will increase impressions outside of ebay, weather could affect it. All those could easily make up a 10 to 15% increase or decrease

Ebay brownouts are another intangible. These are times when they intentionally block some of your listings to some areas

[scotthenrichs]

When you say sell similar, I assume you’re just hitting the sell similar, reposting the exact same item for sale, then cancelling the old listing?

[shrevecity]

When you say sell similar, I assume you’re just hitting the sell similar, reposting the exact same item for sale, then cancelling the old listing?

Yes end the old listing fist then hit sell similar. This is a long discussed thing that has been proven over and over to generate sales. It wipes all views and watchers from that listing and gives it a fresh start which gives a boost in the algorithm. If you cannot list new items daily you should do this with 10 or 15 items a day

[smanzari]

Cool experiment, but its next to impossible to actually gauge - so many external factors to consider besides "Free Shipping"

[Bosoxfan5990]

The biggest takeaway for me is actually the power of the sell similar bump. I knew it was big and thought recycling items to be a worthy use of time, but I'm going to be even more aggressive in the future. Even if I think it's the cheapest, lousiest inventory I've got, I'm not going to let it get past two months. Anything I think that is decent will get recycled within 5-6 weeks at most, and would be considered after one week.

Now that I have the entirety of my inventory listed and I just list new inventory as I purchase it, I have a daily "Sell Similar" process in place.

eBay tells you what items you have that are ending during the current day.
eBay also now allows you to schedule Start Times for your items at no additional charge.

Each morning, I go to my listings that are ending today and manually end them.
Most of my items are lower dollar items, so I don't re-check prices on those. I do spot-check prices on my higher dollar items to make sure I'm still competitive with my pricing.
Then, I go to my Unsold listings, which will contain only the listings I ended that morning. I bulk select the items and choose "Sell Similar." This brings me to the Bulk Editing tool where I can update any pricing for high dollar items, if necessary. This is also where I bulk select the Start Time for all items to re-start them later that day. I select the same time each day.
Once the items have been scheduled, I go back to the Unsold listings tab, select all, and choose Delete from the Actions drop down. This is how I keep track to ensure I'm only re-listing those items that I ended that particular day.

When I'm listing any new inventory, I make sure I schedule the start time for those listings. If it's prior to the current day's start time, I'll schedule it for the current day. If it's after the current day's start time, I'll schedule it for tomorrow.

This ensures my listings/inventory is a month old, maximum, in eBay's eyes.

For reference, I have ~3,700 listings with ~8250 items available.
I do have listings scheduled beginning/ending each day, but every day isn't necessarily evenly balanced with listings.

[cholodolo]

Sales-Why not use sales as the barometer since that's what we're really after? A couple of reasons. Reason #1 is sales have a lot more variance in the short run. Over one week traffic is a better predictor of sales than sales itself. Reason #2 is I did not have a method for combined shipping during the free shipping week. The 30% drop during the free shipping week can be 100% attributable to this. Probably 110-115% in reality.

Takeaways-The first takeaway is anybody trying to claim you'll double or quadruple your sales by offering free shipping is selling you snake oil. I don't know the exact number, but I can say with certainty that it's not a 100% or more bump. I knew that even before running this experiment based on my sales (I have both free shipping and shipping charge listings).


Can you explain this? How can you claim people are selling "snake oil" when you didn't even replicate the conditions, or know the "exact" number?

[shrevecity]

Can you explain this? How can you claim people are selling "snake oil" when you didn't even replicate the conditions, or know the "exact" number?

That's the thing there is one here who claimed without any doubt free shipping DOUBLED his sales within 24 hours.there is zero way that can be proven. There are way too many factors to consider. All this shows is that the claim it doubles anything is likely a stretch.

As many have stated in the past and this does support that to a point that new listings and sell similar listings benefit regardless of if free shipping or not.

If you try a similar experiment your results could greatly differ. There are too many factors that cannot be controlled to give definitive proof.

[prospectorgems]

Thank you for sharing your experiment. I know when you and others were discussing the free shipping in the other thread I wanted to add some of my new listed items with free shipping for the BIN/BO. I found that people were sending offers based on comps of sold prices not including the shipping prices. My counters were actually under the total cost of the item and shipping in recent comps so I don't know if people really take shipping into consideration when making offers. Very small samples size of course but based on what you shared with doing free shipping, I may back those items off and do the shipping charges again and see if that will entice more sales or not.

I can attest that sell similar really does make a difference, free shipping or not. To me though, it's a scary process. Have over 10K listings in my store and I have the categories broken down so it makes it easier, but some categories have in upwards of 2500 listings. For those that don't know how the process works, you need to end all the items and then click the check box at top to select 200 items to sell similar and then continue that process through each page.

The scary part to me is that a few times I had an ebay glitch after I had clicked submit. Fortunately the glitch occurred on a lower item count category so I was able to see which ones didn't actually relist. As far as I know, there are no error reports and since you are doing the "sell similar" it's not considered a relist so the items in the ended section do not recognize as the item being relisted (since it is considered a new listing).

I make sure to take notes when I do the end/relist and I only do it on my computer so I am working from the ended/active items to make sure my number of listings match. If you have a lot of inventory it is can be a nerve wracking long process but it does work.

Thanks again OP for sharing your details, it really does help.

[mogwai316]

Now that I have the entirety of my inventory listed and I just list new inventory as I purchase it, I have a daily "Sell Similar" process in place.

<snip>

Thanks for posting this, it was a really good explanation. I think I'm going to start making that part of my process too, do the sell similar / relist for each listing as they come up on their 1-month expiration date. That keeps it to a manageable number each day instead of trying to do massive amounts at once and taking the risk of losing a bunch of listings to a glitch.

[valleynuckfan]

I've been doing this in an alternative fashion- just ending listings, then re-listing a few days later. It seems to achieve the same goals.

[Eric1285]

Nicely done - really, a great job. I appreciate your analysis and the thought you put into it.

[rwperu34]

Great info. Does "sell similar" work in bulk? (I'm not at a well-functioning computer off my company network often enough to just try it.)

Another advantage to free shipping I remembered this week is on ESE items when you file a claim and don't get the shipping refunded. Seems like you should get that back since, you know, that's the service you paid for that failed.

Yes. You can do 200 at a time.

[rwperu34]

Longhorn had the same question I started with. Did you do these individually, or is there a way to do it all at once? I have been trying to do 10 or so twice a day to refresh my inventory, but its a chore.

Second question: did you change any prices when you relisted them? I have been changing my prices to try to increase sales. Obviously it is working a little bit.

Pick the 10 items you want and end them, then use the bulk listing tool. Once you get there it's one button if you aren't changing anything.

The first week, no. The second week I added the shipping charge into the price, then offered free shipping. Throughout the two weeks I was auditing my inventory, so prices got lowered and cards got delisted.

Lowering price is a very effective way to stimulate sales. You can do this two ways; 1) Run a sale, and 2) Lower the list price. The latter is far more powerful. When the market was still somewhat hot I did this on my graded inventory and saw a huge impact. I also did this in June when I realized the market was going to plop again and it was super effective. Promotional sales are a good way to boost visibility and stimulate multiple item purchases, so be sure and run those every now and again too.

[rwperu34]

Was there any change in multiple item sales?

When you sold multiple items to the same person, did you ship in separate envelopes or combine shipping?

If you combined shipping, did you rebate anything to the customer, or throw in any bonus items?

Yes, huge change in multiple item sales. That's because I didn't have any form of shipping discount. Beauspencer and others can attest to how well buy X get Y% off works with free shipping.

During the free shipping week only one multi item sale. It went in one envelope with no refund. During the shipping charge week there were many combined shipping sales (I charge flat rate shipping). That's why sales are not comparable week to week. It wasn't an apples to apples comparison. It was more like apples to cheesecake.

[rwperu34]

What about intangibles that you cannot control. I can look at my impression charts and see impressions fall later in the month. That is one then you cannot control just number of searches for that item being made. You have a card of joe Bob he throws a ball and hits a kid in the stands searches for him will go way up and will increase impressions outside of ebay, weather could affect it. All those could easily make up a 10 to 15% increase or decrease

Ebay brownouts are another intangible. These are times when they intentionally block some of your listings to some areas

Yes there is variance, but that's why we used traffic. Traffic is very stable once you control for day of the week. The real number might be 5% (meaning I ran hot during Free Shipping Week) or 25% (meaning I ran cold during Free Shipping Week). Even that would be pushing the edge of how much traffic varies week to week (for me).

Right now, this is the best set of data I've ever seen on the topic. It shows two things we can say with certainty; 1) There is a free shipping bump and 2) It's nowhere near 2x.

[rwperu34]

Cool experiment, but its next to impossible to actually gauge - so many external factors to consider besides "Free Shipping"

Yes and no. See previous post. Traffic is very stable. Week one, reset inventory. Week 2 make everything free shipping. Most of what we're catching in week 2 is the free shipping bump. To add to that, week 3 reset inventory again. Traffic drops back to the level of week 1. Perfect!

[rwperu34]

Can you explain this? How can you claim people are selling "snake oil" when you didn't even replicate the conditions, or know the "exact" number?

1) I have a pretty good grasp of variance.

2) This is the best (only) data set I've seen that actually tries to replicate the conditions.

[rwperu34]

Thank you for sharing your experiment. I know when you and others were discussing the free shipping in the other thread I wanted to add some of my new listed items with free shipping for the BIN/BO. I found that people were sending offers based on comps of sold prices not including the shipping prices. My counters were actually under the total cost of the item and shipping in recent comps so I don't know if people really take shipping into consideration when making offers. Very small samples size of course but based on what you shared with doing free shipping, I may back those items off and do the shipping charges again and see if that will entice more sales or not.

I can attest that sell similar really does make a difference, free shipping or not. To me though, it's a scary process. Have over 10K listings in my store and I have the categories broken down so it makes it easier, but some categories have in upwards of 2500 listings. For those that don't know how the process works, you need to end all the items and then click the check box at top to select 200 items to sell similar and then continue that process through each page.

The scary part to me is that a few times I had an ebay glitch after I had clicked submit. Fortunately the glitch occurred on a lower item count category so I was able to see which ones didn't actually relist. As far as I know, there are no error reports and since you are doing the "sell similar" it's not considered a relist so the items in the ended section do not recognize as the item being relisted (since it is considered a new listing).

I make sure to take notes when I do the end/relist and I only do it on my computer so I am working from the ended/active items to make sure my number of listings match. If you have a lot of inventory it is can be a nerve wracking long process but it does work.

Thanks again OP for sharing your details, it really does help.

Sort by oldest inventory, end the listings 200 at a time, then sell similar before you end the next 200. That way if there's a glitch, you're only kinda screwed instead of totally screwed. It would be pretty easy to figure out at that point with just a little bit of time.

FWIW, while auditing my inventory I found about 40 cards that were not listed that I thought were listed. Then I found two that were listed that I no longer had. So stuff happens to all of us. I think that's just something that needs to be factored in. A certain amount of your inventory will get "lost" (the bigger the inventory the more likely this will happen). Just make sure it's $1 cards and not $100 cards.

[smanzari]

Yes and no. See previous post. Traffic is very stable. Week one, reset inventory. Week 2 make everything free shipping. Most of what we're catching in week 2 is the free shipping bump. To add to that, week 3 reset inventory again. Traffic drops back to the level of week 1. Perfect!

There were major differences in the economy, literally a Federal Reserve Decision right in the middle of your experiment (on 9/20) that caused a significant decline to the market (and as a result a noticeable decline in consumer spending), any consistency you're seeing is because of the Algo and liking new listings instead of Free v Charged shipping

[cholodolo]

1) I have a pretty good grasp of variance.

2) This is the best (only) data set I've seen that actually tries to replicate the conditions.

Lol, this is the "grand" experiment?

[thenightman]

Yes end the old listing fist then hit sell similar. This is a long discussed thing that has been proven over and over to generate sales. It wipes all views and watchers from that listing and gives it a fresh start which gives a boost in the algorithm. If you cannot list new items daily you should do this with 10 or 15 items a day

Ever since I read the discussion about this awhile back, I continually end listings at end of day, on the items last day, and then relist using "sell similar" for the next day. EBAy making scheduling listings free has been a game changer.

Thank you for sharing your experiment. I know when you and others were discussing the free shipping in the other thread I wanted to add some of my new listed items with free shipping for the BIN/BO. I found that people were sending offers based on comps of sold prices not including the shipping prices. My counters were actually under the total cost of the item and shipping in recent comps so I don't know if people really take shipping into consideration when making offers. Very small samples size of course but based on what you shared with doing free shipping, I may back those items off and do the shipping charges again and see if that will entice more sales or not.

When I comp new items I always account for shipping costs because in my mind they play a role. If the last 5 sales were (for example):

$3 + free shipping
$0.99 + $4.99 shipping
$3 + $1.00 shipping
$4 + free shipping
$2 + free shipping

I'm putting the total at $17.99 and dividing by 5. That comes to $3.60 on average. If it's a A or B-level player I'll list at $3.99, all others I'll list at $3.49.

[rwperu34]

There were major differences in the economy, literally a Federal Reserve Decision right in the middle of your experiment (on 9/20) that caused a significant decline to the market (and as a result a noticeable decline in consumer spending), any consistency you're seeing is because of the Algo and liking new listings instead of Free v Charged shipping

I mean I literally controlled for this very thing.

The Fed thing raises a legitimate point. That's something I (or anybody) simply can not control for. That could make an argument that the free shipping bump is a little higher than I'm reporting, because it affected the free shipping week and not the other week. Then again, a lot of the Fed actions are priced in before the actual action.

Stocks were down 3% during the Free Shipping Week. BTC is a better indicator for sports cards (IMO) and it was also down about 3% for the week. So call it a 15% gain vs an expectation of a 3% loss. Still a mile away from double. Still an obvious bump. Slightly different number. Same conclusion.

ETA: We also have a 2nd data point after the Fed's decision that is showing a similar bump for free shipping vs with shipping charge. It's not as strong because it's only one day, but it's a point in favor of a free shipping bump. It is indeed a very strong point in favor of the free shipping bump being nowhere near double.

[Budler]

This is not a Grand Experiment.
If you really want to see the different is shipping and free shipping Run an experiment like this:

2 cards of each. One card free shipping the other with shipping. The total price the same. You may want to do a large number of cards. Add them at the same time.

After that you can really see which group sells best.

[rwperu34]

This is not a Grand Experiment.
If you really want to see the different is shipping and free shipping Run an experiment like this:

2 cards of each. One card free shipping the other with shipping. The total price the same. You may want to do a large number of cards. Add them at the same time.

After that you can really see which group sells best.

This would be a great experiment. You could even do it on two, separate, brand new accounts and have quality traffic numbers as well. Unfortunately, I am not in the business of running experiments. I am in the business of selling baseball cards. Doing it this way already set me back a little.

If you see anybody who has run it that way, please, do share the results. Or anybody who has quality data, for that matter. So far the best sets of data I've seen have both been provided by me; this experiment and the results of my graded card sales over the last six months.

[waytoomanycards]

I mean I literally controlled for this very thing.

The Fed thing raises a legitimate point. That's something I (or anybody) simply can not control for. That could make an argument that the free shipping bump is a little higher than I'm reporting, because it affected the free shipping week and not the other week. Then again, a lot of the Fed actions are priced in before the actual action.

Stocks were down 3% during the Free Shipping Week. BTC is a better indicator for sports cards (IMO) and it was also down about 3% for the week. So call it a 15% gain vs an expectation of a 3% loss. Still a mile away from double. Still an obvious bump. Slightly different number. Same conclusion.

ETA: We also have a 2nd data point after the Fed's decision that is showing a similar bump for free shipping vs with shipping charge. It's not as strong because it's only one day, but it's a point in favor of a free shipping bump. It is indeed a very strong point in favor of the free shipping bump being nowhere near double.

The free shipping bump might also have been a "sell similar with edits" bump. Unless I'm confused (very possible), in one scenario you reset your listings without any edits and in the other you edited the price and shipping charges. I'd want to see data when you both reset your inventory and change another variable--such as going back to a shipping charge with an equal decrease in price.

I say that because we know ebay treats both edited and unedited "sell similar" listings pretty much the same with searches that sort by listing date (and I think that's how many card collectors search BIN's. Resetting inventory worked before Cassini). However, we don't know they treat them equally in best match searches.

They can recognize duplicate listings (ie. same title and image), so they could treat the modified and unmodified "sell similar" listings differently in promotions and such if they wanted to. And it would make sense to promote something that a seller edited over something that was identical to an old listing.

[rwperu34]

The free shipping bump might also have been a "sell similar with edits" bump. Unless I'm confused (very possible), in one scenario you reset your listings without any edits and in the other you edited the price and shipping charges. I'd want to see data when you both reset your inventory and change another variable--such as going back to a shipping charge with an equal decrease in price.

I say that because we know ebay treats both edited and unedited "sell similar" listings pretty much the same with searches that sort by listing date (and I think that's how many card collectors search BIN's. Resetting inventory worked before Cassini). However, we don't know they treat them equally in best match searches.

They can recognize duplicate listings (ie. same title and image), so they could treat the modified and unmodified "sell similar" listings differently in promotions and such if they wanted to. And it would make sense to promote something that a seller edited over something that was identical to an old listing.

That's interesting. If I follow, you are saying that because I raised the price and wiped out the shipping charge, that counts as an edit. Whereas the week before I just flat out reset the inventory without edits. So the edit might be contributing to the bump.

Looking at the numbers...it could be! I have the 1st day in week 3, where I put everything back. Those numbers have now been revised. The drop is 9% vs two weeks ago (the original reset) on an inventory that is down about 14% from two weeks ago.

One thing I will say is my normal recycle/reset program does not typically contain any edits and I can still tell I'm getting a bump. The biggest factor I'd say is simply ending the listing and hitting sell similar. An edit could give it a little boost on top of that.

I would also say that I have experienced a big bump in sales from edits too...but the main edit is lowering the price!

[jamcas997]

The Last 7 days I have had 42 shipments. 21 Free Shipping and 21 with shipping. The Free shipping were all ESE and the 21 With S/H were mixed ESE and Ground Adv. With only two days left this month I am about 25% off from last Septembers totals and I had less than half the free shipping items last year. It seems my sales run from January to August and then dump the last 4 months.

[Raleigh504]

Sort by oldest inventory, end the listings 200 at a time, then sell similar before you end the next 200. That way if there's a glitch, you're only kinda screwed instead of totally screwed. It would be pretty easy to figure out at that point with just a little bit of time.

FWIW, while auditing my inventory I found about 40 cards that were not listed that I thought were listed. Then I found two that were listed that I no longer had. So stuff happens to all of us. I think that's just something that needs to be factored in. A certain amount of your inventory will get "lost" (the bigger the inventory the more likely this will happen). Just make sure it's $1 cards and not $100 cards.

I used to do it individually, but the bulk way is SOOO much easier and faster. Can also edit the price (and update the condition since it is requiring it now). This is a game changer for sure.

[shrevecity]

This is not a Grand Experiment.
If you really want to see the different is shipping and free shipping Run an experiment like this:

2 cards of each. One card free shipping the other with shipping. The total price the same. You may want to do a large number of cards. Add them at the same time.

After that you can really see which group sells best.

Catch there is if you use the same account and ebay recognizes it they can nail you for the duplicate item policy or could bury one of the items. The AI is getting smarter when it comes to detecting duplicate items

[Budler]

Catch there is if you use the same account and ebay recognizes it they can nail you for the duplicate item policy or could bury one of the items. The AI is getting smarter when it comes to detecting duplicate items

I'm not into selling on E-Bay. It has been years ago when I sold a card on E-Bay. I do not understand how they work things now-days but reading these postings sure does not appear to be as easy as it was where I was selling.

Yes it appears there are a lot of things that go into the play. If you want to do an experiment like that. Even if one or 2 people worked on a project like this. If you have a card that is on E-Bay with shipping just put yours up for sale with free shipping at the same total price and see if your sells before the other one.

With all of the things that go into play, that can cause an issue with the outcome You cannot really come to a Yes or No answer on. Does free shipping help to sell for the most part. Yes!! some people see it and jump on it, but most buyers look at the total price.

[longhornjunkie]

Does anyone know how to refresh eBay listings via COMC? It seems like if I take an item down on COMC, it just goes "out of stock" on eBay and technically doesn't kill the listing.

[shrevecity]

Does anyone know how to refresh eBay listings via COMC? It seems like if I take an item down on COMC, it just goes "out of stock" on eBay and technically doesn't kill the listing.

That's a comc thing. They would have to do it themselves