Hacker News Reader

Comment by dghlsakjg | Hacker News Reader

There's a grocery chain (Superstore, Canada) which has been using the small two color displays (black white red) for pricing. They use the tiny 2.7" e-paper screens and a TI cc2540 microcontroller to periodically update the pricing via some 2.4ghz signal (ok fine, I was curious so I stole one to take apart). Seems like a VERY expensive way to keep prices up to date. The BOM cost on one of those little guys has to be north of $10, and I would imagine that the vendor that sold them the system doesn't sell them at cost. Multiply that times thousands of products.... So yeah, people are using them, but it doesn't seem like a great system.

View on HN

> Seems like a VERY expensive way to keep prices up to date. The BOM cost on one of those little guys has to be north of $10

That's nothing, even after markup and multiplying it around all the shelves, consider the wages you have to pay for people to go around continuously re-stickering everything. I'm pretty sure it would pay off in under a month and reduce errors... freeing staff to do more useful things. What i'm sure will _not_ be paid off in a month is the cost of integrating it into existing pricing systems.

However I think there is a more practical reason why it's not suited to most supermarkets: they are constantly moving things around due to different products coming and going and different parts shrinking and growing, so why bother with digital labels when you have to keep physically moving them anyway.

I've worked in retail and this is spot on. There are people who did practically nothing but update tags all day. And yeah, any change to the existing software is going to be the real issue. Everything they use is old.

The ones I've seen still need people going around to update tags. They have an NFC/RFID thing that powers that tag while it updates

What is the cost on updating paper label prices? There must be a payback vs. manual labour for these systems.

Typical supermarket with 60,000 SKUS;

$20 total cost per device

= $1.2m to outfit a store.

/$15 employee gross costs per hour

= 38 years of labour for one store....

Either the # SKUs would have to be lower, or the BOM smaller to justify this.

On the flipside consider the material cost of paper pricing tags, reduced risk of pricing error, ability to change pricing more rapidly, increased customer marketing/attraction.

I'd also imagine that removing this function from individual stores and centralising it would save a considerable amount of store management time.

Call me paranoid/cynical but maybe they are planning on a future system were they use the profile they have on you to calculate your preferences and buying power and update the signs on the fly when you're coming up to them.

Or maybe they are already doing that ... I said to call me paranoid.

Some online services are already engaging in price discrimination, though physical retail has greater limitations in communicating prices to individual customers. I think that removing price tags entirely would be necessary for full price discrimination in brick and mortar retail (or some similarly revolutionary step, like AR vision placing price tags on items). That's not to say it's infeasible, but other changes may be necessary first. For example, in an Amazon Go location, you don't need to check out the items at the end. Customers would still definitely want to know prices of items, but I could imagine them caring less about seeing price tags if they aren't checking out immediately after. I think another system that could help would be a partial or full subscription model, where the subscription price could be an opportunity for price discrimination.

You're also forgetting the ability for stores to execute "surge" pricing.

I'm not a fan of it, but that will certainly recoup that cost much faster.

How do you handle the case where a shopper picked up an item at a certain price but by the time they got to the cash register it had changed?

Charge the lower price at the register for 2 hours after changing the display price?

Only change the prices over night? (but unlike with employees you can reprice the entire store every night without any particular cost)

easy, you just display both prices with a fat title 'Hurry up! Price will be 25% higher in 12 minutes"

If this doesn't exist already, it will soon.

This seems like a bad idea. If I own a physical store and you’re in it I don’t want to encourage you to leave quickly. I want you in there as long as possible putting things in your cart.

wouldn't this encourage people to buy fewer items?

But customers are so used to confusing pricing labels you could just make the old (cheaper) price say: "Discount: $x".

Also if you hook up the change to motion detectors where you know no one is in the aisle, then you can skip it altogether.

Could maybe be solved by adding some more technology (e.g. a barcode) that allows them to reserve the item at the current price, plus maybe some other kind of incentive.

The customer probably won't notice and if they do they'll walk back to the place they found the item and find the e-ink price tag has a price that matches the till. If they some how dispute it further the business will just lie like a rug until a particularly pernicious customer takes pictures of the price tag and then argues with the manager.

The eventual outcome will possibly be a class action law suit and in a few years the customers can sign up to receive a gift card.

You can see similar behavior with the same company here: https://www.cbc.ca/news/business/loblaws-25-gift-card-class-...

Keep in mind that shelves must be restocked as well. Employees are putting their hands on most product's shelf space daily. Updating the prices while restocking seems like an almost negligible cost.

That said, I know some people that work in the grocery industry, and my impression is that the margins are tight enough that they do the math on EVERYTHING

In terms of their hardware cost and software development/wireless network, it's much more than just Superstore. Superstore is indeed huge but it's just one part of a much larger gargantuan grocery conglomerate, Loblaws.

https://en.wikipedia.org/wiki/Loblaws

they operate with other brand names for non-superstore-sized regular grocery stores, and the T&T markets which are their own market niche: https://en.wikipedia.org/wiki/T_%26_T_Supermarket

From observing the behavior of my local superstore it also appears they're using the eink price tags on shelf in conjunction with some sort of software for planning shelf space and inventory, so it has some value to them other than just the manual labor to change paper price tags.

Hey do you work in EE/Hardware? If so, would love to send you an email about a project I’m working on

I am not the OP but I do. I'd love to use small color e-ink displays for low power indicator displays on equipment I design. I only run into it with hand held battery stuff, so for stuff I plug in it's kind of a non-issue.

Feel free to visit my website and send an email if you think you've got something that we could work together to get into some of my clients' projects. If it makes the final product more compelling I'm definitely open to it, especially if it runs on trivial drivers that a low power microcontroller can handle.

I'd basically expect a display that works with a chip like a STM32WB (wireless integrated so makes sense to pair with a low power design). https://www.st.com/content/st_com/en/products/microcontrolle...

Of course fancy people have access to Qualcomm parts, so I understand if that's what you're aiming for. But if it's suitable for a bit less fancy of microcontroller chips and has a reasonable price I'd love to replace the literal LCD displays we're still stuck on because I guess it's 1996 in the realm of equipment prototyping.

I do as a side hustle/ passion project. I make a small single use board, about as complicated as an arduino.

But I am by no means an EE.

Do you have a link to your project page?

$10 BoM and the replacement costs attributed to curious nerds :)

Reply on news.ycombinator.com