2012.24 | Retail Robots

Robotic Store Staff – Carnegie Mellon has developed a robot for retailers to assist in validating stock outs and misplaced items in the store.  The autonomous device rolls through the store on its own and scans the shelves to validate locations of items and notify staff via iPads.  They have augmented that solution with some additional technology – a Google Streetview like view of the store showing where products are located on a large format digital sign.   The sign is augmented with product information should customers wish to view it – even trailers of DVDs on the shelves if they are interested.

This is a very ambitious undertaking and that team should be commended for their initiative.  There are a few hurdles for this team I can see with this type of solution based on my experience in retail:

  • I don’t know how the system works as far as where the robot moves in the store.  If it isn’t bumping into walls like my Roomba, or if the Kinect can’t recognize obstacles, it probably has a map of the store in memory somewhere.  Updating store layouts and planograms is a lot of work, and generally where wayfinding falls down.  I’m sure it’s relatively straightforward in one CMU bookstore that probably doesn’t change out their mugs and sweatshirts too much in any given cycle.   Try this in a specialty or apparel chain across hundreds of sites with varying floor plans and the potential to move store fixtures, and it becomes much more challenging.  There would need to be a tool to accommodate tweaks at sites to administer this to validate that the store information.  The challenge will be around local versus remote administration.  Local staff know the store layout but are probably not technical enough to update the map on the robot.  Remote staff can update the robot but won’t know the store.    Even better, let the robot figure it out autonomously – that’s the ultimate.
  • The Google Streetview kiosk layout is very interesting, but once again, stores and merchandise are constantly updated.  I’m finding Google Streetview is already getting out of date – the stores on the street have changed since the images were taken.  I don’t always trust it now.  Same goes for a store.  The product changes, the store changes.  There needs to be a constant update mechanism.  Even if the product is shown in the right place on the Streetview interface, users will think that the view is different and become confused.  It’s the updates that kill solutions like this.
  • That User Interface on the digital sign had a lot going on.  I’m sure a great deal of thought went into it, and it looked great, but it has to be so easy that my mom can use it in front of a half dozen spectators with some product under her arm.  Keep what the solution does as simple as possible.  It’s not an app on a mobile device for users with time on their hands and lots of buttons to push.
  • Finding a product is more of an art than a science.  That’s why people are preferred to machines so far.  Describing product is harder than it sounds.  Examples on product search are always something easy like a CMU mug.  When a customer comes to search for something in the store, it could be a specific brand and easy to search on.  More often, it might be that lavender shirt with the grey buttons – do you have it in size 4?  While customers will walk up to a screen with no other options, they will prefer to deal with a real person.  It’s always easier to walk up to someone and ask.  Voice activated or image based search and validation would be terrific – ideally it could ask the customer some questions to narrow down the items and then show pictures of the product to validate what the customer is looking for, and then where it is in the store.  If not available, allow online order and ship.  Another option?  Let the robot provide directions to the product. If they could walk faster, having the robot lead there would be incredible.
  • iPad notifications are useful, but without followup, it doesn’t mean anything.  The solution should ensure that staff are notified and prompted to action.  If no update is made, then there should be automatic escalations to management.  Would also be great if the solution would indicate if product that should be on the shelf is in stock in the store or not.  The system could prompt re-ordering for outages.

All technology solutions have challenges. Tying the solutions into the operations without impacting store staff’s ability to get their jobs done is what will make or break any retail technology solution.   This is a very interesting idea with lots of potential.  I hope it gets built  out and there is interest from retailers.

Restaurant Robots – A restaurant in China has opened with robotic staff.  The robots actually usher in customers, cook food and serve.  They are also anthropomorphic and candy coloured to impress the children.  Given the cost of the robots, and the fact that this is likely not an automat – but requires human intervention, I would expect that the food is relatively expensive.  Biggest unanswered question – what do you tip a robot?

Robot Model for Fit – Purchasing clothing online can be convenient, but fit is always a factor.  Sizes can vary widely by clothing brand.  For those shopping for clothing online, the practice of purchasing a couple of sizes and just returning the ones that don’t fit at no charge has become common.  While a convenience to the shopper, it is a rising overhead cost for retailers.  In an attempt to reduce returns based on fit, fits.me, an Estonian company has developed a robot which can change its shape to fit clothing, and then provide measurements across thousands of points for each garment.  When a customer enters their measurements, and selects a garment, it shows how the item would fit their body.  This could allow online shoppers to have a better idea of how clothing will look and fit on their body – ideally reducing returns.

Android Salesperson/Actress - In order to push some extra sales for Valentine’s Day, Takishimiya’s Tokyo store turned to a real android.  Last February, an eerily realistic robot sat in a display case using her android phone and passing the time.   Straddling the line between mannequin and real life model, the robot reacts to its surroundings to provide lifelike responses to those around her.   While a novelty at present, you could see a hyper-realistic android catching on as a way to fully demonstrate products – show how easy it is to move and stretch in athletic apparel perhaps.

Other Important Retail Robots to Remember:

Kiva Systems – Amazon owned Inventory Picking Robot system – link from 2009 Post .

PAL robots – Used in Abu Dhabi Mall.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 678 other followers

%d bloggers like this: