Hacker News new | past | comments | ask | show | jobs | submit login
Gap acquires 3D fitting room startup Drapr (retaildive.com)
78 points by vitabenes 88 days ago | hide | past | favorite | 66 comments



Personally I don't think the issue is "3D fitting" - the issue is that the sizing and fit of clothes hasn't both been: standardized and quantified. Do both and a 3D fitting room wouldn't be necessary to begin with.


> Personally I don't think the issue is "3D fitting" - the issue is that the sizing and fit of clothes hasn't both been: standardized and quantified.

Yes, it has. The problem is that nontrivial clothing requires multidimensional fit which is impractical for mass-produced off-the-rack clothing, and marketing of mass-produced clothes is therefore driven by all kinds of things, none of which is using the well-established standardization and quantification of fit, which is used only by people making bespoke clothing or altering off-the-rack clothes for specific individuals.

Of course, unless it is paired by on-demand alteration, 3D fitting doesn't really address that problem.

But its not meant to (though the for-the-public image may be that), its to deal with the problem that taking clothes to actual fitting rooms increases handling, damage, theft losses, and the resulting policing and remerchandising is a major source of labor requirements for shops.

For the Gap, the problem this hopes to solve, or at least mitigate, is “retail employees”.


As technology advances, it becomes easier for us to standardize and quantize the fit of clothes in increasing dimensions. And so instead of one product having a certain cut and then it’s just got a few sizes, we could have a continuously variable design which could be produced in 10,000 sizes. Stores could carry popular sizes but any prospective buyer could check if they have their ideal fit range in stock in any items, or could order the correct size by mail.

Of course the garment industry is still heavily reliant on human labor for manufacture and the model I describe would really only work if the process is basically fully automated. But it would be nice if my clothes didn’t rely on marginalized labor in Bangladesh so I’d like a fully automated source.

Anyway you’re not wrong that it’s impractical with the current way of doing things, but it seems fully automated clothing production is advancing, so perhaps in 20 years it will be possible.


>Yes, it has.

It most certainly has not. Even in men's legwear where your clothing is more or less fully defined by two measurements (waist circumference and inseam length), there are lots of variations between brands for garments with the same nominal size.

I have 2 pairs of billabong board shorts that are slightly too big and fall down if I don't cinch the drawstrings extra tight. Size 34 waist.

I have another 2 pairs of o'neill shorts that fit perfectly and don't fall down even if I don't lace up the drawstring at all. Size 34 waist.

Let's not even get into women's clothing where you can have such ill-defined terms like "size zero".


> It most certainly has not.

It has, but, again, mass market clothing doesn’t use it. Bespoke clothing and tailoring does.

> Even in men's legwear where your clothing is more or less fully defined by two measurements (waist circumference and inseam length),

No, those are what mass-market panta tend to use as nominal size dimensions, with variations in the relation of those and other standard dimensions being brand or style variations (where there are options for some of them within a brand, typically multiple of them are collapsed together under another single axis like slim fit/classic fit/loose fit, which tends to lump ankle)

The standard measurements for men’s pants include all of:

waist circumference front and back crotch length hip circumference waist to knee length knee circumference thigh circumference calf circumference ankle circumference waist to ankle length inseam

(sometimes, some of these will be estimated based on visual inspection and the other measurements.)

> there are lots of variations between brands for garments with the same nominal size.

Yes, mass market clothing nominal sizes have no consistent relationship to even the things the subset of measures they noninally represent.

> Let's not even get into women's clothing where you can have such ill-defined terms like "size zero".

The only difference between women’s dress sizes and men’s pants sizes is that the former is a one-dimensional arbitrary ordinal category and the latter is two dimensional, and pretends to a (false) correspondence to a standard length measurement.


>Let's not even get into women's clothing where you can have such ill-defined terms like "size zero".

size zero at least makes sense if it's the smallest, and they're using zero based indexing. It really gets ridiculous when they have size 00.


I've started selecting for brands that have more information than those two measurements. Seat and taper are more helpful, for instance.

More casual button-downs are also very frustrating, as sleeve length and chest circumference aren't provided accurately for S vs M vs L


> Size 34 waist.

If you want an extra shock: measure your waist. It's not 34 inches!


Agreed, and that's also not taking into account vanity sizing, where the differences in sizes is on purpose


One problem with that is that not all the articles are the same size. You can order 1000 t-shirts in the same size from a manufacturer and not all of them will be the same size (especially if you order the cheapest you can find).

This is mainly due to cheap fabrics stretching unevenly (roll to roll difference and beginning of the roll can stretch differently from the end) and if you cut them by pressing through a thick stack the ones at different spots in the stack stretch differently during the cut and thus end up different size even if "cut" to the same size.


What if we ended up with a system where I know I'm a B29.

Then after every shirt made, it's tested against the B26 - B34 (since it's probably somewhere in there) and then labelled as a fit for B29.

Then I just go shopping for B29s.


You'd be mapping a multi-dimensional space to a single dimensional one, which might work, but would probably be less intuitive. In the very least, we'd probably want something like a small/medium/large, then some extra info about the shoulders, and then a little bit more info for guys who want a tight fit (those guys with abs to show off). You could map those all on to 26-34, but I think it would not really help much.


This is also a male / female -type body distinction.

I'd argue men have about 4 primary body shapes (roughly, height:shoulders:waist:hips), within which most are +/- a bit.

Women have... a lot more.


Yeah. I implicitly assumed we were talking about guys because the idea of simplifying women's sizes just seems intractable.


"Sorry, B29 is a mild outlier in terms of sizing - only 5% of the population wears B29 so we don't keep that in stock."

I had this problem when I was younger with an outlier waist/inseam ratio. I was only one inseam size out of "normal" but had to catalog order my pants. I wonder how it would work now with modern supply chain efficiencies. Back then I could rely on a warehouse somewhere having my size, but just-in-time may not allow room for that.

I love the idea though. As someone who rarely finds a t-shirt where I like the fit (either too tight or too long), I'd love to just have something I could count on. I can't even rely on just getting the same size from the same store each time.


If we got away from factory clothes and moved to a model where someone made two or three year's worth of clothes for $500 (well I guess the price would be based on your packages??)

Then the clothing designer would input B29 and it would adjust all sorts of things and print out a pattern. Then the clothing designer could "test fit" the final shirt on a really cool adjustomatic mannequin made of weird pistons and balloons. Ok that last part is unlikely to happen but it sounds awesome.


> If we got away from factory clothes and moved to a model where someone made two or three year's worth of clothes for $500

Lovely idea but what do the poor fools do whose weights are going up and down? I mean I know that's not 80% of the population on a given year but there is some value to being able to buy just a few garments at a time that fit approximately correctly right now, even if they won't really in another six months.


Maybe Etsy, or re-seller sites that have used clothing.


Assuming B29 is a series of measurements, this is the right solution in my opinion. The challenge is the cost of measuring and tagging every item after production. The relationship to model out is cost of pre-measuring every garment vs cost of returns. If you could dramatically lower your return rate you may be able to be more profitable measuring every garment. But... even measuring successfully presents challenges.

Going in to a store would seem absolutely insane if you could just order tons of stuff in your size and it all fits and you would not have to make returns.


As someone totally ignorant of this industry... how can those problems be fixed? Is it an issue of using higher quality materials, or some factory process change?


Inexpensive, quality of construction, reliable sizing.

Pick two.


Body types vary widely though. If you’re athletic you may have a strong shoulder to hip drop, getting shirts that are blousy in the midsection. Conversely, if you’re heavy set or short you may get garments that are overly long.

Some companies do try to address this to a degree with slim or athletic cuts, and yes a tailor could take you the rest of the way ‘mostly’ there, but the future of clothing is 3D scanning and printing.


As exemplified by the old threadbase post: https://web.archive.org/web/20160318074519/threadbase.com/un...

The sizes vary within brands themselves.

Traditionally the solution was to get a tailor take your measurements and make a pattern for you and cut from cloth to make you a garment. Obviously, in today's day and age, that gets a bit expensive for most people.


There are companies working on scaling this kind of thing, though:

https://www.indochino.com/suits/fit

That one is a "measure at home" affair, but I've also heard tell of ones where there's a pop up at a co-working space or whatever, and you go get measured, pick your fabrics, and get the shirts in the mail a few weeks later.

Cost-wise, it's definitely more than getting them at a department store, but it's not ridiculously more. Think like 2-3x rather than 10x.


I'm waiting the day when I go into a store and the fitting room scans my body (naked or just underwear) and then it creates an outfit just for my body. We could probably pull it off already, but it can't compete with third world produced on price, and those are good enough.


I think there have been failed attempts at this over the past 20 years or so. If you're willing to wait, they could do this somewhat affordably --just won't be off the rack fast (week or two?)

That said, a tailor will make you pose properly and not slouch or droop your shoulder, so they would take better measurements. Presumably, SW could correct for bad posture as well.


Measurement is easy. Even the elastic deformation (i.e. multiple poses).

Manufacturing at scale is currently impossible.

Tailoring is basically robotics' worst nightmare: input material with varying characteristics from batch to batch, delicate materials, multiple different motions and tasks, a non-reducible movement and orientation space (e.g. sew a hem along a curve), and a complex mapping between physical tasks and end result (different stitch on seam = different motion).

We'll get there (and if it weren't for low cost labor, would already be there for t-shirts). But it's really hard.


I can't imagine 3D printing clothing working for most kinds of garments any time soon. Just mechanically stitching straight seams with a robot has proven to be nearly impossible. It might be economically feasible to laser cut some fabrics and drive down MTM costs, but you can already do that fairly cheaply with existing tech.


A 3D fitting room isn’t going to solve the issues you mention tho.

I also think what you’re describing could be easily solved with more measurements of clothing you’re considering.


I’ve done just this, with a tailor doing tours for a MTM business. Took about 15 minutes to do 18 measurements. With those in their system, I can custom order any number of (dress) shirts customized in a variety of ways and they do fit well.

The issue is it just doesn’t scale well abd is limited by types of garments on offer, no standardized way to share measurements with other manufacturers, and requires we redo the process or for me to ‘guess’ at adjustments if my body changes in any material ways - they only come to major cities once a quarter or so.


3D fitting is first step for custom fitted to order clothes, like Amazons "Made for You" t-shirts https://www.businessinsider.com/amazon-custom-fit-t-shirts-u...


> Body types vary widely though.

The GP is more referring to companies who deliberately don't conform to standards. Best example of this is women's clothing that alter sizes so they are more appealing to women who don't traditionally fit that size number.

TL;DR - "Omg I finally fit in a size 2 dress" is a huge selling point, even if its not categorically true.


Understood, that’s an issue in itself.

But as someone who is on the shorter side with an athletic body type, almost no clothing produced today (aside from high end made in US brands) fits me at all. Vintage stuff does. No amount of standardization will get around that. I have to pay more for tailoring, getting custom made, or spending tons of time searching out boutique brands that work for my body type.

What I’m suggesting fixes it all.


A recent 99% Invisible dove into the history of standardized sizing that was really interesting (apparently most of the data on human measurements used in industry comes from US military research): https://99percentinvisible.org/episode/on-average/


Sizing is pretty quantified. A tailor has specific measurements they make. (Presumably women's clothing is the same.) One problem is that you end up with a huge number of permutations, no one is "average"[0], people have different preferences when it comes to the type of fit, etc.

This sort of thing has been being talked about for 20+ years. The problems include that you're probably never going to get to the equivalent of a tailored suit--even after taking all their measurements, tailors often will make some final adjustments. On the other hand, most people are fine with their day to day polo shirts, T-shirts, and trousers being off the rack without paying double/triple or whatever for customization.

[0] https://99percentinvisible.org/episode/on-average/


> Sizing is pretty quantified.

Sort of, but it's all fake if you don't have the right QC strictness at manufacture. People like to joke about "fake size numbers" and "S/M/L" and "different style cuts", but forget about that and focus on waist and inseam measurements for a moment.

Men's pants have been labeled with waist and inseam measurements in inches forever, and, differences person to person and brand to brand entirely aside, the same brand of pants in the same style in the same color with the same size label on the same day can have different actual waist circumferences by more than half an inch so that one pair will fit _you_ and another won't fit _you_ because their assembly tolerance is higher than your fit tolerance.


>being off the rack without paying double/triple or whatever for customization.

I bought a suit a while back - they tailored it (pants and jacket) for <$20. This is "really expensive" if it is a $10 t-shirt, but not that bad if it is an $80 point-down shirt, and even reasonable if it is a $200 pant/jacket combo


That seems cheap for tailoring in the US. I don't really wear suits or even blazers any longer but, when I did, I'd either get them custom-made in Asia or extensively tailored in the US. Haven't done that for ages though.

I was mostly referring to day-to-day wear that's $10-40 or so off the rack.


> I'd either get them custom-made in Asia

I'm selfishly disappointed China's takeover of Hong Kong because it gives me very mixed feeling about getting suits from there. They do amazing work, but supporting them is, in a sense, supporting China's takeover, even if the shop had nothing to do with it.


Sizing isn't really quantified at all. Brands use vague terms like "Small, Medium, and Large" that aren't even consistent within the brand, and depend on the cut and particular style.


Mass-produced clothing sizing is certainly inexact. But the body measurements for an individual are reasonably well specified.


What does “body measurements for an individual are specified” mean? Like body measurements on the size charts are specified? Those don’t tell one of the clothing will fit, only that there’s a possibility it will fit. And in reality the ranges they give are far too wide.

And one can specify their body measurements all they like, but it barely helps selecting a size online or even in store.


Actually one of the issues is also standardization.

Finding someone that is average on just four or five dimensions is extremely rare. Bucketing people into a handful of clusters and taking those averages is only a marginal improvement. Brands standardizing to those same clusters make it impossible for many people to find fitting clothes.

It’s just like shoes. For a given length, at most about 30-40% of the foot widths can be accommodated without offering multiple widths. And most shoes only come in one width.


This is especially true for Gap. Their sizing is bonkers compared to the norm. I stopped shopping there a long time ago because it was too much of a hassle to go and try things on in person. Instead, I order online from retailers that have consistent sizing and that I know fit me well.


Congrats guys!

For those who don’t know, Drapr was a YC co from last summer’s batch: https://www.ycombinator.com/companies/drapr


Zozo solved this with the Zozosuit: https://corp.zozo.com/en/news/20201029-6375/ (that is version 2).

You wear a ridiculous looking figure-hugging suit and scan yourself with their app. Then you don't wear the suit again (unless you change shape). Version 2 of the suit looks much less silly.

Now you can order what you want and it's for your size!

It does works, but... make sure you know exactly what you want. e.g. do you want a closer fit? Looser fit?


What problem/s does 3D fitting room solve?

What problem/s does 3D fitting room introduce?


In theory, it works to bridge the gap between an in-person visit to a clothes store and an online shopping experience like Amazon.

When you buy clothes online, you can only see images of the clothes and see them on other people. You can't really tell how they'd look and fit on you.

When you buy clothes in a store, you can try them on but you're left with the selection of a physical store (e.g. it might not have your size, you might want a different colour).

With a virtual fitting room, you can, at least in theory, get the best of both worlds: a large selection of items and the ability to see how they fit on you.

After going shopping with my wife recently, I came up with a dream for how the experience of clothes shopping would be in the future and a virtual fitting room fits into that vision:

- You digitize your body (e.g. by visiting a store with a body scanner)

- You browse a catalog, on a computer or phone at home or perhaps on some kind of AR mirror in a store (this is where the virtual fitting room fits in)

- You have the items you like the most delivered either to your home or to the store

- You try them on for a final check

- You keep the ones you like and return the ones you don't (either by post if you're at home or just hand them back if you're in-store)

Depending on the specifications available to the retailer, the virtual fitting could also do analysis of the garment and your body to tell you how good a fit it is. The CAD packages available for patternmaking have stress analysis which can show where the garment is too tight or too loose and that could be displayed to the customer based on their own measurements.

The same concept could also be used to eliminate sizes from the customer's mind. The store could use the customer's measurements and the technical details of the garment to automatically supply the best-fitting size.


Presumably what is my number one issue with online clothes shopping, and why I have and never will do it - sizing and perception.

1) Sizing - ordering multiple sizes so you get the right one is not a sustainable or environmentally friendly way of purchasing clothes online.

2) Perception - looking at a good-looking, slim person trying on clothes is not representative of most people. I've seen plenty of clothes look good on a mannequin look awful on me. Wish I could say vice-versa, but the mannequins don't talk back.


Even sizing can be 'odd' sometimes. I can many times grab 2 different items the only difference supposedly is color. One will fit the other will have weird spots that dont. I basically have to try everything on before I buy it. Otherwise I am just going to have to return it immediately. That is currently easier in a store than playing the mail it back game.

If I were ordering custom sized cloths then maybe this would be useful. Neat for sure. But I am not seeing the usefulness because of the wild inconsistencies in most cloths.


On 2) perception. Not only are the models typically in the top 1% of fitness fit with 40” chests and 31” waists and six packs, but it’s rare they’re under 5”11 despite the median male height being 5’9”.

So all in all the male clothes models usually look like just a fraction of a fraction of the population and that fraction is all outliers in no way representative of a typical shopper.


> is not a sustainable or environmentally friendly

These products need to be shipped to the U.S., usually by boat, where freight trucks haul them to warehouses before another set of freight trucks deliver product to storefronts which have giant parking lots next to them so customers can drive their gas cars to park in them. Let's not talk about all the unsold waste.

I'm open to options.


What does the back end look like?

Is every garment custom made, or selected from inventory?


I've no idea. I've never bought clothing online and don't trust these services either (yet).


The problem I’ve seen with all the online fit systems is they tell you the closest size even if it’s a horrible fit, they never tell you the brand’s sizes aren’t good for you.


> What problem/s does 3D fitting room introduce?

Nasty PII leaks


> by showing them how an item will actually look on their body

This is only half of the equation, and I find a lot of people are ignoring - mainly because it seems intractable.

As important as how an item of clothing looks on my body, is how it feels on my body. Does it constrict when I move. Am I able to sit, stand, and walk comfortably in it?

Some times a piece of clothing can look perfect, if I am standing rigidly like a mannequin, but it is highly uncomfortable with any movement.


No amounts mentioned. Can I assume it was a pittance then?


If they’re an acquisition target this early in their life then it must be because of potential, not inability to raise money, as being a part of YC would make it straight forward to raise capital to continue to identify business opportunities (if they were struggling). I’d be very surprised if it was under 20m.


Crunchbase says they only raised $125k so even if they were acquired for a few million, they likely made out okay.


That seems to be the YC batch amount.


Owning this data myself seems worthwhile. Being able to query a shop's clothes without giving up my dataset, i.e. each measurement and each shirt being anonymized so the vendor can't simply see all my measurements. Think there's a name for this process.


This seems to be an acquihire? IF you wanted to get a high valuation you would have stuck it out by yourself longer I would think?


Gap, Inc. also has a collection of dying brands and is a little desperate. Gap and Banana Republic are stale. I guess Old Navy is doing OK. Google trends says Athleta seems to do OK, but it's a second-rate Lululemon.


I worked at a big retail company a few years that was desperate to make investments in tech. Probably didn't help that HQ was in Silicon Valley. They ended acquiring a 3D company for a similar purpose to Gap only the company was actually a total mismatch for what they intended. They produced 3D models, but not actual usable software. The tech leadership internally advised against the acquisition but the CEO did it anyway. Then we had to pivot them to basically build what we wanted from scratch with no clear market need. AFAIK, it's been languishing for a few years.


There's a lot of interesting discussion here!

I am the ex-CTO and co-founder of a company that does 3D-scan-based fit, specifically ML-based morphology extraction from 3D data of people (www.treedys.com , not a YC company ;-) ).

A lot of the points raised here in the comments are very valid: there is absolutely no consensus on standardisation of "fit"... but even if there was: "fit" is a very complicated concept that means different things to people in different settings. Furthermore there is often no consensus even inside big clothing companies as to which grading system (this is the "fit catalog" so to speak) is to be used for a specific collection / year / style etc... and those companies that do have this kind of thing standardised are often fighting with internal staff that grew up in a non-digital era to keep the documentation up to date, make sure the right files are shared, push people towards using digital design tools consistently etc...

But this does not mean - by any stretch - that the current system cannot be improved! There are many things that can be done with morphological data that are not immediately obvious in terms of improving fit for individual end users. Comparing "tshirt measurements" VS "human measurements" is the obvious starting point, but this is not actually the best way to make real-life recommendations, and there are many better systems that are possible once you have high quality metrics! Here is one for example: what about completely disregarding the "extracted body measurements in inches", and just building a flywheel where you take people's entire set of say 250 measurements as a vector, and use this data point along with a set of items they have bought (for themselves) and not returned. Once you have this kind of data for, say, 5000 people and 50 SKUs of items of clothing... you can build a very granular recommendation system, based on morphology, which will take into account many aspects of "fit" that are not immediately obvious in the measurements...

One thing is for sure: the big incumbents in fashion retail have felt the sting of shifting to an online-first world. In 2020 the total dollar amount of merchandise returned from online retail doubled (https://nrf.com/media-center/press-releases/428-billion-merc...). Literally: double the volume in one year. This was especially impactful in online fashion, where returns have always been a giant cost and a limiting factor to expansion ... but up until 2020 fashion retailers had been delaying action while desperately trying to modernise other aspects of their businesses, and morphology data felt like "too big of a problem for now.

Now, in 2021, with a few big players like Amazon taking the lead in "morphology-based fit" the situation is changing... and even the slowest fashion brands are starting to realise that not having a fit / morphology strategy in 2021 is starting to look like an automotive manufacturer NOT having an EV strategy in 2012 when Tesla was releasing the model S...

We're feeling it: after hanging on by the skin of our teeth for 6 years building deep tech solutions ( 3D ML to map a naked body on messy point cloud data of a clothed person for example, that was a tough nut to crack!), we're suddenly getting a lot more inbound from people we could not get on phone for a 5-minute conversation 2 years ago. And we finally have our first public-facing deployments in-store, using the full capture-to-ML stack (all developed in-house) that we have been crafting all this time! It feels good to finally get a product out and start getting the feedback loop going after so long.

All of this to say: congrats to Drapr, They made a great product and their timing was excellent! Let's hope this helps GAP achieve it's sustainability goals and reduce returns and waste, because in the end this will be the major benefit to all of us :) But in any case: morphology-based-fit is now "an idea whose time has come", and we can expect a lot of cool new things to come from this over the next few years!




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: