1

Last month, 3D Hubs communicated to its suppliers via the Changelog and via email that it was going to carry out a US Express Checkout experiment in order to simplify the checkout experience for customers. As a result of the experiment, 3D Hubs has decided to provide customers with the option of Automatic Hub Selection. Customers can still select a Hub manually if they wish to. Below is some more information on the experiment and the team is continuing to innovate and test in order to improve the customer experience and benefit the community as a whole.

Why is 3D Hubs running the US Express Checkout experiment?

There are two main reasons:

  1. The number of prints ordered has doubled over the last year and the number of Hubs has grown even faster. To improve the reliability of the platform for customers, 3D Hubs prefers to direct them to Hubs that perform well according to our performance metrics captured in the ā€œHub Scoreā€. The Hub Score includes metrics such as customer ratings, average response time, decline rate of orders and several more. Customers do still have the option to select Hubs manually should they wish.
  2. Our customer research has shown two key things. Firstly, that customers want certainty when it comes to their part quality. Secondly, the ordering process can sometimes be overwhelming for customers due to the number of options presented to them. Customers need to consider the printability of their models, determine which material to print with and then decide which Hub to use to manufacture their parts. By fully separating the material choice and automating the Hub selection, weā€™re hoping to simplify the customer experience and eventually also standardise the part quality. The Express Checkout experiment has already shown a positive impact on the amount of orders now arriving at Hubs.

Which Hubs are selected for Express Checkout?

For the majority of the Express experiments, 50 American Hubs have been pre-selected based on Hub activity level, Hub Score, the materials they offer and if they have a business registration (as this is often required by our professional customers). We also give preference to Hubs that offer flat-rate shipping and have ā€œinstant orderā€ enabled as this supports a simple Express Checkout experience and increases customer satisfaction.

As this is an experiment, these criteria will continue to evolve and the number of Hubs will probably increase based on the data that we collect. If a Hub has been left out of the initial Express Checkout experiment, this may change as::

  • We optimise our Express selection criteria
  • We expand the total # of Express Hubs to keep up with demand
  • Your Hub Score improves
  • Your Material offering expands

How do Hubs get selected for order fulfilment once they are included in Express Checkout?

Over the last week, weā€™ve run several experiments with this to measure impact on conversion:

  • Picking the highest ranking Hub based on Hub Score
  • Picking the best value Hub based on Hub Score + price
  • Distributing orders among the top few ranked Hubs for each request

In the long run, we want to factor in a combination of price, Hub Score and location to make sure customers get high quality, fast and cost-effectively.

One of the initial issues with Express was that the customer location wasnā€™t properly captured. This resulted in many orders going to just a few Hubs as the distance was no longer factored in. This issue is currently being resolved and is the number one priority for our development team.

Why did some Hubs started getting less orders?

A key reason is that part of our customer base now has the option to select Express Checkout and currently t only 50 Hubs have been selected for this as we continue to test and improve the service. Another reason is related to settings in the new material selection flow which are currently being updated:

  • The new material selection is clearly better understood by our customers (15% increase in conversion) but in the initial, current version niche materials are not shown. Weā€™re currently working on a database migration that will allow Hubs and customers to select from all 3D printing materials on the lowest, brand level. This should go live next month.
  • By default the most popular color was already selected for each material. Hubs not offering this color did not show up in the listings. The default color setting will now be changed to ā€œno preferenceā€
  • By default the most popular layer height was selected for each material. Hubs not offering this layer height did not show up in the listing. Weā€™ll now change this so that Hubs offering a lower layer height will also show up.

Why is there a waiting list for Hubs?

Due to the reasons outlined above, the number of Hub signups is growing so fast we cannot provide enough orders for all Hubs signing up. Since July, weā€™ve only be onboarding new Hubs where we have a shortage of supply for certain geographies or materials.

Where did 3D Hubs communicate about these Express Checkout changes?

Weā€™ve communicated these changes twice on the 3D Hubs Changelog and via email to all Hubs selected for Express.

The combination of iterating fast and keeping our Hub base happy is a delicate balancing act. We realise that the Express Checkout experiment could have been communicated better and apologise for any confusion. In the future, weā€™ll ensure that all big experiments and transitions are communicated not only via our Changelog but also by an email to all Hubs. They are key contributors to our ecosystem. With great Hubs as our partners, we can provide customers with a consistent, high-quality and professional experience to help us in our mission to decentralize manufacturing worldwide.

26 Likes
2

Thanks for explaining this better. My other gripe that I didnā€™t see mentioned is that as a hub we have no idea how our hub is comparing to others or how we are ranked. There is no metrics available to us anymore except our score such as 4.96/5.00

I would also like to see an option on the changelog to sign up for emails when there is a change added there.

20 Likes
3

So, making my customers order from a hub in Fargo ND is better than letting them order locally? So forcing them to pay for shipping and waiting longer is better than being able to pick up locally and pay no shipping?

Got it.

7 Likes
4

Thank you for the thorough explanation. I would just like to see the local hubs shown automatically below or beside the autoselect.

Also if you know there are these bugs in the autoselect it doesnt make for the most accurate stats so why dont you take it back down for the meantime until you fix the bugs?

6 Likes
5

Well, itā€™s good to know the little hubs donā€™t matter. You guys take take you BS and lies and shove it. Iā€™m done.

7 Likes
6

Hi ABram,

First of all, thank you for the explanation. Would have been better though if this explanation had been given upfront instead of after many complaints.

But some things are still not entirely clear for me.

I do not know how many Hubs there are in the USA, but do you mean that right now all customers who are part of the experiment in the US are now redirected to just 50 Hubs?

Have these customers been informed that they are now part of this experiment? If so, do they have an option to opt out?

That would basically mean that, right now, 50 Hubs are getting the orders that would normally be spread among many times more Hubs all over the US.

Is this correct?

If I understand your stats from the Top Cities listing, New York alone already has something like 300 Hubs listed.

Does the Autoselect option now means that just a few Hubs in that NY area are now receiving the orders that used to be divided among those 300 Hubs.

4 Likes
7

This is my understanding as of today. I find it highly disturbing that they SHUT OUT ALL BUT 50 HUBS. You guys arenā€™t getting order because only the ā€˜chosenā€™ 50 are. This is a new revelation and not in line with what weā€™ve been told this far. They Lied to us. There is no more doubt. There is no other explanation. Iā€™m furious and will be taking action to tell everyone in the 3d printing community I know to stay well away from this unethical site. Iā€™m looking for the replacement. Somebody message me if you know of something so I can tell others. 3dhubs, you guys are subhuman assholes. Way to ruin a good thing. Good riddance.

4 Likes
8

Pretty sure this is only effecting 15 pct of orders as part of the experiment.

2 Likes
9

Where did you get that 15% of orders from?

I specifically do not see that mentioned in the original post.

And even if ā€˜onlyā€™ 15% of all the orders are now being directed to just 50 Hubs, that still means a lot of Hubs being left out and will not receive any more orders or will at least see a large drop in order rate.

While at the same time, those 50 ā€˜chosenā€™ Hubs somehow must (already) have the capacity for large volume printing.

2 Likes
10

I remember seeing it an email atleast for the initial experiment, although the Changelog does say 33% for the first one. Its a bit unclear what the percentage is for the v2 of the experiment.

11

@Anyoneat3DHubs Did you all switch how Express Checkout it works again? I was never notified that I was selected as a hub for express checkout yet as long as I make my hub the least expensive in the area (NYC) it chooses my hub as the hub for express checkoutā€¦as soon as I raise it above the next cheapest hub it goes to them instead. So thatā€™s my first question, is it now just choosing a hub for express checkout based on price? And second, is my hub actually one of the hubs chosen for Express Checkout or is this just some sort of bug? (and if I am now part of Express Checkout why wasnā€™t I notified as you mentioned you notified the initial 50?) Thanks.

2 Likes
12

ABram thanks for taking the time to post this information. I really appreciate it. I could tell from several past orders with various customers that the ordering side of the service was problematic. Most of my customers either did not care about the layer height or did not understand it (Whatā€™s a micron?). Particularly when prototyping, the color does not matter and in some cases the type of filament does not matter either. These customers also asked basic questions about 3D printers and how it works but were still willing to use your service. I live in the south east and everyone has a million dollar idea! (I have discovered!). A lot of these customers were cosplay designers too.

For the more technical customersā€¦they chose my hub because they were looking for something specificā€¦ninjaflex, .150 layer height with large build volume, design work, etc. So it may also be about the initial landing page on 3dHubs and designing it to cater to two different types of customers - someone with a great idea and technical designer/engineer.

I look forward to being able to reach customers again through the optimized search methods your development team delivers! Take care! Hope this helps!

5 Likes
13

Thanks everyone for their feedback. Iā€™ve listed your questions and provided my answers below:

Most of my customers either did not care about the layer height or did not understand it (Whatā€™s a micron?). Particularly when prototyping, the color does not matter

Yep, good points. Thatā€™s why by default the color has now been set to ā€œno preferenceā€ and weā€™ve pre-selected the most popular layer height for the chosen material.

There is no metrics available to us anymore

Indeed the initial version that we launched a few months ago had several bugs so we decided to temporarily take it offline, but the new version will go live in the next two weeks. My apologies for the temporary inconvenience.

I would like to see an option on the changelog to sign up for emails when there is a change added there.

The tool that we use for our Changelog is Headwayapp, and they donā€™t allow for that unfortunately. But going forward for changes with large impact we will make sure to email in advance.

I would just like to see the local hubs shown automatically below or beside the autoselect.

Thanks for this idea, we will consider this as we continue to develop and improve the Express Checkout.

So it may also be about the initial landing page on 3dHubs and designing it to cater to two different types of customers

Weā€™re trying to capture this in an updated material selection design in Q4. People who are new to 3D printing should be able to easily select the right material from a popular pre-selection and engineers should be able to sort and filter in the the hundreds of different material options.

I was never notified that I was selected as a hub for express checkout

Hubs selected for the Express Checkout did receive an email. As part of one of the experiments a few Hubs outside the original selection of 50 could have received an Express order. In the next two weeks Hubs should see a notification on the order page if the order came in through automated instead of manual selection.

Right now all customers who are part of the experiment in the US are now redirected to just 50 Hubs?
Have these customers been informed that they are now part of this experiment? If so, do they have an option to opt out?

The US customers that select the Express Checkout option are indeed directed to one of 50 Hubs (one exception in the experiment, see above). Customers still see a button to manually select a Hub so that I consider the opt-out.

Its a bit unclear what the percentage is for the v2 of the experiment.

In the V2 version 100% of US customers will see the Express checkout, with the option of manually selecting the Hub.

Also if you know there are these bugs in the autoselect it doesnā€™t make for the most accurate stats

Most bugs have been fixed. Resolving the niche materials being excluded will take a bit more time as this requires a migration as a few months ago weā€™ve started creating a completely new material database. This migration will also enable Hubs to indicate their material offering on a much more granular level.

Does the Autoselect option now means that just a few Hubs in that NY area are now receiving the orders that used to be divided among those 300 Hubs.

Indeed for orders that come through the Express option and not the manual Hub selection, they are distributed amongst a smaller group of Hubs. In reality because of our performance based ranking not all of the 300 Hubs in New York were actually getting regular orders.

5 Likes
14

"Its a bit unclear what the percentage is for the v2 of the experiment.

In the V2 version 100% of US customers will see the Express checkout, with the option of manually selecting the Hub."

Why donā€™t you have a ā€œBasic Modeā€ which uses your Express Checkout, and an ā€œAdvanced Modeā€ where is has the old interface. These modes would be selectable in their user profile with basic being the default. Currently, it is now more difficult for an advanced user to locate filaments like IGUS materials for printing linear bushings. Advanced users are currently being forced to use a limited radio button/pulldown list for finding advanced materials which often do not fit these categories.

Currently, we are in Starbucks trying to order a lunch sandwich Bistro Box but online ordering is making us navigate the drink menus to somehow get to sandwiches. I donā€™t even want to figure out how to look for something specific (like layer height etc. hehe)

15

Hey Bram,

Thanks for the reply. I was wondering about how this new express check out will effect repeat customers? A large portion of my customers come back on a repeat basis. I imagine their workflow is searching the Boston area and then finding my hub again since its on the 1st page. With the express check out that gets significantly harder to do. Is there potentially a way for customers to save hubs they like in the express check out flow to then re-order from? I apologize if a feature like this already exists, I havenā€™t personally ordered a 3D print on the network in a while.

1 Like
16

Soā€¦ basicallyā€¦ what your sayingā€¦

You have shut down all but 50 hubs in the USA and funneled all the orders in the USA to those select 50 hubs.
Yeaā€¦ thatā€™s not at all underhandedā€¦ nopeā€¦ *sarcasm*

Would have been nice to inform your hubs in the USA va ****EMAIL**** that they will no longer be getting orders thru normal means on your platform.

And you canā€™t say you ****DIDNā€™T**** shut down all the 3dhubs in the USA except those 50 because you can choose any of themā€¦ your Little button you added at the bottom of the page to allow the customer to look for them has been cleverly colored to blend in with the page and easily overlooked. I even missed it myself looking for it.

IF you plan to shift to servicing only your ā€œProfessionalā€ customersā€¦ you need to be telling your hubs that they are no longer needed and they can take their services else where instead of just hiding them all from general customers view.

P.S.
ā€œit was in our changelogā€ is a HORRIBLE excuse. Something this MAJOR should have been emailed.
Plain and simple. no ands ifs buts. DO IT!

I award you no points.

9 Likes
17

v2 clearly says 100%

2 Likes
18

Actually to expand on thisā€¦
I fully expect to see a EMAIL notice sent out to your active USA Hubs detailing these changes as most will have missed this post and frankly even the hate threads on the internet. Your unwillingness to notify your hubs directly even after the fact is very concerning from a professional and business standpoint.

5 Likes
19

Give all of your customers a direct hub link, or a link to a page with your embeddable order widget. Both can be seen via Tools and Perks.

20

Hi ABram,

Thank for the clarification.

I need to think about some of your statement, but I would very much like to make a suggestion right now.

There a growing number of materials available which outperform the traditional standard PLA or ABS materials usually chosen.

For example, just recently a (Dutch) company launched a ABS-variant which outperforms standard ABS in basically every mechanical parametric you can think of. Personally, I increasingly encourage customers to try one of these improved filaments because they usually do not only deliver better performing parts (like parts that can be sterilized at 134 degrees Centrigrade for medical applications), but also because they are often easier to print with.

Why not make it simple by grouping materials together in basic groups. A customer can select the basic group of materials he is interested in (like ABS or PLA) and after than can go into more detail if necessesary?

This way, no hub would be left out (a major complaint it seems) and a customer should ALWAYS be able to find a Hub to print his/her/their project.

1 Like