Where Is The Material Pricing?

Discussion in 'General Discussion' started by ChristianH, Apr 21, 2018.

  1. MrNibbles
    MrNibbles Well-Known Member
    Getting back to pricing methods... I came up with a set of models that when plotted in various ways help to reveal what factors are being used to price prints. Again useful at a service that doesn't publish exact methods or rates. Of course this is somewhat dependent on a service to provide their measured values of material volume, machine space, and surface area after model upload. If they don't you'll need to calculate those yourself - with the machine space potentially being the most difficult to calculate depending on your test model geometries.

    I generated two basic sets of models that hold either material volume or machine space (volume) constant and vary the other. The first set is a cube hollowed out by various sizes of smaller cubes with a small through hole from the outside to any inside cavity. Therefore this set varies material volume while holding machine space constant.

    The second set is another cube spread out into various shapes to vary machine space, but the material volumes are constant for each shape. The size of this cube (37mm) was chosen to have approximately the same material volume as the most hollowed out cube of the other set to provide an intersection point.

    Here are the models used. (The spread4 model is hollowed from one end. The hollowing not observable from the preview thumbnail.)

    pricing test models.jpg

    I generated plots on a spreadsheet for material price as a function of material volume, machine space, and surface area. The idea is that holding some model parameters constant will reveal pricing slopes on certain graphs. Basically a straight line with a slope reveals a cost dependency on that parameter. First up wsf price vs material volume and wsf price versus machine space graphs. We can see that there's a price relation of $0.28 per cm^3 of material and $0.21 per cm^3 of machine space.

    pricing charts wsf.jpg

    Plotting wsf price versus surface area gives a plot that shows no perfectly straight relation between surface area and price (there are slight curvatures). In fact the two model sets show different correlation trends - one positive and one negative. I haven't decided yet if it would help to generate another set of models that holds surface area constant and varies material volume or machine space. So far I'm not sure that's necessary or easy to do.

    pricing charts wsf 2.jpg
    A fly in the ointment is determining any constant part handling fee or minimum baseline fee that transitions to material/machine/surface area costs. For determining those you need models that are closer to minimum size to show those pricing transitions. The method outlined here is primarily good for determining various individual linear price components although some additional mathematical manipulations might be useful to tease out those other cost additions. Another fly in the ointment is price schedules changing for oversized objects.

    I suppose another way to do all of this is to create a wide spread of models where all the basic model parameters have large variations between models. Then generate a basic equation that includes a bunch of pricing methods and run a Monte Carlo analysis to fit the pricing data of all the models (for a specific material) to an equation. I'm not there yet.

    Oh, and one more mistake I made was that the 37mm box I chose is too small for ceramic models (minimum would be 40mm on the 2 large bounding box sides). I'll probably adjust those, rinse, and repeat the whole shebang.
    Last edited: May 15, 2018
    chanusa and ChristianH like this.
  2. Anyuta3D
    Anyuta3D Well-Known Member
    Did you SW people changed the names of materials? I just noticed that you renamed the FUD material from "Frosted Ultra Detail" to "Smooth plastic" and the FXD from "Frosted Xtreme Detail" to "Smoothest plastic". So now we have the 'Smooth' and the 'Smoothest', right? (reminds me THIS)...

    Btw, re-naming materials does not solve the problems eg the well-known frosting issue on FUD parts.

    PS 1. Changes also done to WSF too (now re-named as "Versatile")?
    PS 2. The end of May must be the period of year where everybody in SW is free to experiment.
    FerretDesigns and crashtestdummy like this.
  3. MichaelAtOz
    MichaelAtOz Well-Known Member
  4. ChristianH
    ChristianH Well-Known Member
  5. ChristianH
    ChristianH Well-Known Member
    Still nothing new from SW.

    So, what's up SW?
  6. MrNibbles
    MrNibbles Well-Known Member
    I made the files public for use at Tinkercad if anyone wants to try a similar spreadsheet approach to reverse-engineering prices for specific materials at any vendor. If you're really smart you could probably automate the process using the API, spreadsheet download from Shapeways, or other web thingees that are beyond by ability. (As far as I can tell pricing has not changed from earlier in the year.) The 9 files I used are on the large side and could all be scaled down by the same factor so you're not dealing with such extremely high prices. Porcelain is gone so the minimum sized object can be smaller than what was used across all materials. Here's the link to the public files at Tinkercad. If you have an account there you can copy them over to your user space and edit them there for size or whatever. You'll also need a Tinkercad account if you simply want to download them to your computer.

    ChristianH likes this.
  7. MrNibbles
    MrNibbles Well-Known Member
    I had some brief moments of free time and decided to extract some material pricing for Shapeways and another service whose name shall not be spoken. Just wanted to see what problems might pop up when doing this. It's kind of easy to do this for Shapeways since I don't think any price schedules have changed intentionally and I still have a good feeling for how they price various materials. But the problem needs to be approached as if you don't know anything and just want to use bench marking files to figure out how prices are calculated.

    Here are some of the problems:
    1) Figuring out equivalent material names across vendors. Vendors will make it a point to give materials their own unique names. If it's not a market distinguisher its a market confuser. White SLS nylon at one place might be called super duper premium natural laser beam fused plastic at another.
    2) Altered process parameters for competitive materials. Maybe you're company W and offer the same material as companies X, Y, and Z. Company W can distinguish their version of the material by offering it using slightly different layer thicknesses or a proprietary post processing procedure. It's not always easy to compare apples and apples.
    3) Pricing results displayed on multiple pages or tabs to thwart automated data collection techniques, or at least make things a bigger pain in the butt for inquisitive people (or competitors) to do it manually.
    4) Don't explain exactly what parameters are being used to price a material. For example material volume, machine space, and surface area might be the most common but there could also be others such as bounding box. Don't make it clear which parameters are used in combination for any particular material. Don't post any of the calculated parameters (material volume, machine space, etc.) for users to plug into a spreadsheet for analysis. :)
    5) Large pricing discontinuities that may not become evident using bench marking files. For example price changes for vastly different sizes of models, critical hole sizes that open up machine space price reductions, or special handling charges for super large and bulky models.
    6) Are pricing parameters linear? It helps if they are linear if only to identify them visually on a graph. However it's not a given even though at this point that's probably what most of the industry does. Any pricing formula can become as complicated as a vendor might feel it needs to be to guarantee profit.
    7) Different shipping prices. I'm not considering different shipping prices but there can be large differences between vendors.

    So given all of these problems I only looked at Shapeways wsf, er, uh, white natural something or other and the white unpolished nylon from another vendor, aka vendor Censored. The spread sheet data and analysis charts are shown below. I took some liberties such as using Shapeways calculated parameters for material volumes, surface area, and machine space even though that may not even be a thing at vendor Censored. The prices for the two models with the most material volume did qualify for a cheaper "economy" shipping discount at vendor Censored so the analysis would need to be extended using larger models to figure out what's happening for larger models. However even these discounted prices were more than the base Shapeways price.

    I think at least another class of models needs to be added to better figure out models that use surface area for pricing. In this set I use a couple of tiny models to help figure out minimum pricing, a group of models that has constant machine volume but varies material volume, and another group that has constant material volume but varies machine volume. The idea behind this is that you want a group of models that exercises a specific pricing parameter without affecting the other parameters. So for things like polished cast metals that may include a surface area adjustment we would need to add another group that varies surface area but holds machine and material volumes constant. Things can quickly get out of hand with too many pricing parameters!

    Also note that I scaled down the files mentioned in a previous post that are available at Tinkercad by a factor of two to get more reasonable and commonly sized models. You can redo this analysis using models all scaled by the same factor.

    What happens after you think you've extracted a pricing formula for a model? You really need to back verify it to see how well it matches to reality. I haven't done this for these simple examples but it's easy to add a column for expected pricing on a the spreadsheet.

    Anyway, here are the extraction results for the white strong flexible natural nylon something something something for Shapeways and vendor Censored.

    wsf pricing extraction data censored.jpg

    wsf pricing extraction censored.jpg
  8. lawrencekramer2014
    lawrencekramer2014 Well-Known Member
    Thank You! Great to have this analysis.
    The biggest differences between vendors I am seeing in silver and gold is in the initial setup charges, vs reruns.
    If the vendor is able to make a mold and use that to build to waxes of subsequent runs, instead of additive methods,
    the subsequent price per piece can be a small fraction of additive.
    Metals cost is most of the cost, and that does not vary much.
    Some models work for molds, some do not.
  9. MrNibbles
    MrNibbles Well-Known Member
    It's quite a pain in the butt doing this, particularly for multiple materials. It would be easier if there was a pricing analysis club where one person would volunteer to thoroughly examine 1 specific material's pricing and keep up with it across multiple print vendors over time. Ideally there's be a spare HAL9000 sitting around somewhere willing to do the daily price check updating and spreadsheet analysis when prices change.
  10. MrNibbles
    MrNibbles Well-Known Member
    And now for regular sandstone. Analysis is easy for Shapeways but still a relative unknown for vendor Censored. Part of the problem is there possibly being several parameters that factor into the price. My guess is that adding in bounding box information (which is kind of sort of like Shapeways machine space ) will get things closer. But this illustrates the problem of not having pricing formula transparency.

    sandstone pricing extraction censored.jpg

    pricing extraction data censored.jpg

    eta: I calculated that the bounding box volume appears to have a price impact of $0.25/cm^3. However I still can't find any combination of parameters in a pricing equation that gives a close approximation to actual cost in sandstone. I also haven't been able to find any info on the web about how this vendor might be calculating the print price. So annoying!
    Last edited: Aug 21, 2018
    Ngineer and ChristianH like this.
  11. ChristianH
    ChristianH Well-Known Member
    Great stuff!!
    Please PM me about what is under the 'Censored' label.
  12. MrNibbles
    MrNibbles Well-Known Member
    My goal is not to sway people towards a competitor but you could probably figure it out for yourself. :) I'm simply trying to give tips as how to extract pricing schedules from vendors that aren't very transparent about it, which now appears to be just about everyone. It's my experience that other vendors are generally more expensive across the board for most materials although some might be cheaper for certain size ranges.

    Quite frankly I selected Shapeways as my primary 3D printing point of contact because of good transparency and having the best store front options - even if a lot of changes over time have not been rock solid with decent backwards compatibility. I get the feeling my near term efforts would be better served by archiving everything in a safe place and revisiting opportunities somewhere, somehow in a few years. I might be persuaded to play around with a Spring and Wonder type thing with a few of my better selling and more popular models but other than that my store, which relies heavily on CustomMaker, is so kaput as to not being worth maintaining or adding product to right now. So in the meantime its all about generating better price extraction model sets and spreadsheets!
  13. DoctorOctoroc
    DoctorOctoroc Well-Known Member
    I don't even bother responding to messages anymore lol. It's mostly people asking for different scales of existing models and I can't bring myself to spend the 10 min or so it takes to upload, scale, list, etc. the model if there's even a chance the pricing on it has changed from what it used to be. Maybe some people haven't seen any changes there but for me, every new model I've uploaded since the changes has something inexplicable (to me, currently) that makes the price higher than it should be - mainly when it comes to the dyed plastics. I still can't figure out why some models have the flat $1.00 fee and others have a random number higher and proportional to the size of the model. If it's a glitch they're taking their sweet-ass time fixing it.
  14. stannum
    stannum Well-Known Member
    Volume, machine space, file history, etc gone in new pages? Or moved somewhere else?

    BTW, the new pages are spreading, even if you aren't in the beta group.
  15. reducedAircraftFactory
    reducedAircraftFactory Well-Known Member
    It will be interesting to see this same sort of analysis applied to the new pricing model.

    From what I've seen (in very limited experiments), the new minimum price is $12.50 for unpolished, white Versatile and Fine Detail plastics, whether Smooth or Smoothest. Perhaps the floor is less with something *very* small -- I should try a 1cm cube to be sure. As you approach some cutoff point (bounding box would be my guess), it jumps to a $15.00 minimum. I would guess there are a set of "N" floor prices driven by something simple like bounding box or maximum linear dimension, and no matter how small your model is, you can't get it cheaper than that floor price.

    As things get larger, a different set of criteria take over, such as material volume and other parameters, because the prices no longer conform to a nicely-rounded "floor price" and you get numbers like $43.01 instead.

    I have only looked at a handful of models thus far, so it's too soon to draw curves or firm conclusions.

    Here, again, is a place where a little pricing guidance would be golden. Can we get two small models in under that $12.50 floor? What if they're linked or sprued? Does it matter how they're arranged? Since Shapeways won't tell us the rules, we have to learn through experimentation. We have become pricing alchemists.
  16. MrNibbles
    MrNibbles Well-Known Member
    I'm taking some time to better archive my old designs. Most are in Tinkercad and I'm changing name conventions there, uploading them to my hard drive, and then re-uploading them to Shapeways so everything is consistent with the same model name. I'm trying to get this done before any pricing changes so I can be more confident I'm actually consolidating around the same exact model, and in some cases I may have rescaled the design once it got to Shapeways.

    As far as reverse engineering pricing goes that might need a rethinking of the methodology. If they are changing prices on an hourly, daily, weekly, or monthly basis I'm not even sure what benefit there is to having a price calculation formula and it becomes a huge pain to do that unless the process can be better automated. It might make more sense if there was a maximum base price and they only reduced prices to encourage additional sales when the factory is being under-utilized, but if prices are going up and down willy-nilly that would be simply crazy.

    The best part of all of this, however, is the apparent death knell for that previous goofy Shapeways marketing theory about prices needing to end in 0.99 cents, that setting a higher mark-up can make your product appear more exclusive and therefore more attractive to buyers, and hopefully the end of the final price setting mechanism to set your markup amount.
    Last edited: Aug 29, 2018
  17. crashtestdummy
    crashtestdummy Well-Known Member
    For me the whole process as added a new layer of complexity to my buying and not in a good way. Now it comes down to:
    1.do I need this
    2. do I need this today or gamble with the pricing moving favorably, kind of like playing the stack market but with even less information
    3. if its too expensive, do I wait, give up or suck it up and pay.

    Given my stuff is currently ether for my own use or purchased from me directly, I wonder if anyone else has those kinds of thoughts about purchasing?
  18. ChristianH
    ChristianH Well-Known Member
    I know it's not of any help at all but SW just needs to know.

    ShapeWays. Your new scheme sucks!
  19. reducedAircraftFactory
    reducedAircraftFactory Well-Known Member
    Let's experiment with the new pricing scheme for White, unpolished and undyed, Versatile Plastic (nee WSF).
    I've got a 1:144 two-seater reconnaissance scale model that was $7.06 under the old pricing (no markup, no shipping...just the raw model price). That will be our reference, and with a modest margin it used to see fair demand.

    Under the new pricing, that same plane is $15.00. There won't be many sales at that price, sorry. It is 212% of the old price and it is beyond most customers' price-tolerance.
    Hmm, $15.00 is a suspiciously round number. (Good thing I didn't try this exercise in Euros!) It sounds like I'm hitting a minimum floor price based on bounding box size or maximum dimension. Let's try a 1cm cube... that comes out to $5.00 -- again suspiciously rounded.

    Well, if I'm going to hit a minimum, let's see if I can slip another identical plane in there for the same price. (Note that the models do not touch.)
    Almost -- the price for the pair is $17.50 rather than $15.00 for the single plane. That's a better deal for the customer, at $8.75 per plane.

    There used to be a "per-part" charge in some materials. Let's try stacking them together and linking them by a sprue, so that they come out as a single part. (The single sprue is too weak for this to be a good design, but it's just for experimentation.)
    That worked -- the price for the pair is only $15.90 (down from $17.50), or $7.95 per plane. Not bad, and better than the side-by-side.

    Maybe the old trick of linking the two models together with a removable ring will work (like a chain)? In theory, that should make them a "one part model".
    Hmm, no. That model costs $17.50, the same as the product without the chain-link. So the algorithm is either not charging per part or it can't figure out a chain-structure. Or maybe... since the bounding box of the stacked planes is smaller? No, it couldn't be that poorly done. Let's try the same two planes with just more empty space in between.
    Sure enough, the price is now $24.65. What? It's not smart enough to figure out it can pack separate parts into a smaller bounding box? If that's true, it's going to make a big difference to my price based on how I pack separate parts into a bounding box.

    What happens if I pack the two planes together into the smallest possible bounding box without them touching?
    $15.00! Wow, that's the same price as a single plane. And at $7.50 per plane, that's not drastically more than the old single-plane price of $7.06, though it forces the customer to buy two when they may have only wanted one. And the generated model picture is a mess -- it would be hard for a customer to tell what they're looking at.

    So the good news is that -- under the new pricing model -- I can continue offering models at a reasonable per-plane price if I want to rework almost every model I offer. (I do not count that as a good use of a weekend.) Plus, communicating that to customers will be a challenge: "with this material, you get two planes; with that material, you get one."

    The fact that the pricing software cannot seem to figure out how to auto-pack multipart models and eliminate white space is hard to understand.

    On a side note, the pricing of Fine Detail plastic seems to be directly related to the amount of plastic used, with some extra charge for support material (just like the old days). And the pricing of Professional Plastic is just not worth commenting on -- Shapeways is not pricing that material in a way to encourage sales (IMHO).

    I look forward to seeing others' experiments with the new pricing model.
  20. ChristianH
    ChristianH Well-Known Member
    Great writeup reducedAircraft.
    Once again it's confirmed that SW lost the ball and that both designers and customers are getting screwed over.

    SW, are you even reading this? Caring? Guess not.
    MikeyBugs95 likes this.