Insight-Driven Retailing Blog

In 2004 Google left the comforts of search and advertising to try its hand at web-based email. The success of that product was followed by Google Talk, an instant messenger, and Google Docs, a web-based word processor and spreadsheet. In 2008 Google released its Chrome browser, at which point it had matched Microsoft's major products, except one. So today Google announced its upcoming Chrome OS, a Linux-based operating system for PCs.

The initial target is netbooks, and the OS will be very lean, relying mostly on the browser to access applications. Others can decided if its possible to completely abandon Microsoft and rely solely on Google products. My questions is, can this stack work for a point-of-sale?

If you talk to any retail chain of substantial size, they'll tell you that at least one store is offline at any given time -- more likely its a handful. That's just the nature of wide-area-networks today, and while I'm sure it will continue to improve, it will never be perfect. So its paramount that a POS be able to sell while offline, and relying on a browser is risky.

I realize there exists point-of-sale products that do just that: they are web-based and rely on being online. They make use of AJAX and unique peripheral drivers to make it work. If you're in a retail business where you need a connection to activate products, like wireless providers, then that could be a fit. But if you're "scanning and bagging," then some resiliency is necessary.

I suppose Google Gears addresses this concern, but I haven't seen a POS based on Gears yet. If you know of one, let me know. It would be interesting to understand that approach.

So from the retail perspective, I think the forthcoming Google Chrome OS is just another Linux distribution in an already crowded market. Because the Google name is attached means its positioned for success, but not necessarily for cash registers.

As IHL analyst Lee Holman puts it, "most consumers have adapted to self-service as a way of life." Its the only way I checkout at Home Depot, and I see kiosks by ZoomSystems in all the major airports now. There are Red Boxes at most McDonald's allowing people to rent DVDs, and when's the last time someone pumped your gas for you (NJ residents excluded)?

According to Infinity Research, "the retail industry is the largest employer of self service kiosks; one third of all kiosks produced are implemented in the retail industries, which is also the fastest growing market for self service kiosks." RIS News reports that sales from kiosks will surpass $775 Billion in 2009.

The latest kiosk to hit the retail market is Best Buy's "game trade-in" kiosk. It accepts modern console games (PS2, PS3, Wii, Xbox, and Xbox 360 titles) and provides store credit for the trade-in, assuming the disk passes its quality tests. Eventually you may be able to buy and rent games and movies from the kiosk as well. Alternatively, you can get an estimate of the value of your trade-in via Best Buy's website, then mail the game to Best Buy in exchange for a gift card.

There's a more in-depth review of the kiosk over at JoyStiq.

I find that self-service machines can't replace competent employees, but there are so few of those anyway; so I'm willing to trade the hit-or-miss customer service for the predictability of a machine.

Software as a Service, or SaaS for short, get's lots of attention. I've even discussed it here on this blog in the past (Retail in the Clouds and To Host or Not to Host). And why not? Its an interesting topic. But I want to point out there's another side to story, and I think its value may be even greater. Data as a Service (DaaS) is not a new concept, just a new label. Companies have been buying data for years -- the internet just makes delivery of that data more timely.

In the retail industry I can think of a few obvious examples. Retailers often buy tax rules from companies like ADP/Taxware and Vertex (two active partipants of ARTS). These are updated often, so monthly feeds are necessary. Retailers, especially grocers, often buy consumer market research from companies like Neilsen and IRI. This data is used to determine product roles within category management.

Consumers have been benefiting from DaaS as well. Shopping sites like Shopzilla, price comparison sites like BizRate, and coupon collectors like MyGroceryDeals have pulled data from various sources, sliced and diced, then provide the data as useful information. Sometimes the consumer pays for this directly, but often its the retailer and/or manufacturer. This field is getting more exciting as data becomes more accessible via mobile devices.

The third area of DaaS is data integration and enrichment in the cloud. For example, companies like StrikeIron upload your data, enrich it with external data, and deliver it back over the internet. They are ofloading the data managment concerns of the enterprise. Companies like Kognitio are even providing data warehousing priced by the amount of analytics run. (I'd still be nervous about outsourcing my companies most valuable asset, but then I'm risk adverse in this area.)

So next time someone starts talking about SaaS, ask about its older brother DaaS.

Getting all or most of your software from a single vendor has several advantages. First, the acquisition costs are lower because you don't have to go through the RFP processes repeatedly, and you can negotiate a volume discount. Second, integration costs are typically minimized since the software components have been pre-tested to work with each other. Third, it's often easier to find a single implementation partner that is familiar with all the constituent pieces and therefore has a lower price. And finally, the support burden is lessened since there's always just one vendor to contact.

Retailers understand that best-of-breed often brings with it complexity and complexity is the enemy of IT. Building a strategic partnership with a software vendor can be a win-win proposition. Gander Mountain is another example of a retailer that's gone down this path. As Pitou Keo, Senior Director of Information Services at Gander Mountain puts it:

"Oracle is a valued partner and we are pleased to build on our strategic relationship with them. We expect to achieve a number of significant, short-term benefits from our investment in the Oracle Retail solutions while also establishing a strong foundation to support our long-term growth."

But building such a strong relationship takes time and trust. Luckily I've been seeing more of these recently, and so has Joe Skorupa over at RIS News. According to a blog posting, he's seen the 10 point spread between software selection philosophies (integrated solutions suites vs best-of-breed) increase to 16 points during the latest round of surveys.

I guess if its easier to get all you groceries from one store, it makes sense for software as well.

When I first heard about credit card number tokenization, I couldn't believe I missed such a simple solution. I was so focused on complex encryption that I didn't "think out of the box." Since then I know of several retailers that selected that solution, and several others that have chosen to stay with encryption.

So what exactly is tokenization? I came across an interesting article in which Shift4's Randy Carr uses a princess analogy to explain the concept:

"Say you have a castle with a princess, and all these bad guys keep riding up trying to kidnap her. The way the industry has approached security is to put a moat around the castle, bar the doors and windows and put archers on the roof. Why don't we just remove her from the castle?'"

Once an authorization is received, a token is generated that replaces the actual credit card number. The token follows the constructs of credit card numbers, so it meets validation requirements in existing applications. From an application's point-of-view, the token looks just like a credit card number.

For all subsequent interactions with the acquirer, the "token middleman" converts tokens to credit card numbers. This allows the retailer to claim no credit card numbers are stored in its databases, yet still be able to process charge-backs and returns. The security onus is back on the gateway, off the retailer's premises.

Companies such as BluePay, EPX and MerchantLink are using this approach to help retailers meet PCI compliance.

But how is this different than encrypting the credit card number? Couldn't a token simply be a secure hash of a credit card number? Yes it could, but the elegance of the token is that it mimics the format of credit card numbers so that existing applications work as is.

Of course when a card is swiped, the credit card number needs to be protected until it can be tokenized, so some encryption is still necessary. But that's transient encryption, so keys are temporary and require very little management.

As Shift4 says, "They can't steal what you don't have."

Best Buy's Remix API (described in a previous posting), seems to be catching on with developers. There are now several applications displayed in their sample gallery. I think they've proved that if you built it, they will come. Now I wonder if its really driving sales?

Tesco, the global grocery chain headquartered in the UK, is getting on the bandwagon with TescoAPI. Like Remix, the web services expose product and inventory information in the hopes that others will create innovative applications and widgets that will drive additional sales. It appears that Tesco has taken the Microsoft Azure route, but we won't hold that against them. The original API was done using SOAP, but apparently they are transitioning to REST soon.

I love this trend toward openness because it fosters innovation, and that can lead to better shopping experiences for consumers.

You can get access to over 50 APIs categorized as "shopping" over at Programmable Web. They range from the Yahoo's (added in 2005) to Tesco's (added this week).

Yesterday Best Buy launched a new service called Twelpforce that is aimed at using store and GeekSquad staff to answer product and technology questions via Twitter. According to Bloomberg, they will even be doing some TV commercials on the topic. So I decided to see how this works.

The first thing I noticed was that most of the Tweets were of the "hello out there" variety, but that may just be because the service is so new. It looked like there were tweets from both store associates as well as GeekSquad technicians. There's no rhyme or reason to the account names. Some clearly represent stores (e.g. BBY1094), GeekSquad technicians (e.g. Agent 13700), specialties (e.g. CameraJenn), and individuals (Ingandela). BestBuy doesn't seem to be limiting their creativity.

So let's see if I can get an answer. Below is my first conversation:

I kind of figured that the real-time nature of Twitter would have gotten me a faster response than an hour. Nonetheless, the response was accurate and helpful. In fact both responses were good, but I wonder if there are in controls in place for limiting the number of responses. I could see getting deluged by multiple responses, and in this case both gave the same information.

So then I asked a follow-on question, but there's no good way to associate the related threads.

After two hours there was still no response.

For me at least, the Best Buy forum is actually better because I can have a conversation that's captured in threads. Plus when necessary, I can send a private message that contains more specific purchase information when necessary.

The premise here is that a person has a problem and knows to send his/her question to Twelpforce. In that case, I would guess the customer is probably already a BestBuy customer. But what about the customer that just needs help. Isn't that a great opportunity for BestBuy to find and capture the customer? My next experiment:

It seems to me that software could comb Twitter for opportunities to help people with their electronics. This would be a great way to establish a relationship with a potential customer. BestBuy even recommends their staff do this, but my question went unanswered. I have to assume either Twelpforce is not searching on "plasma TV" or can't distinguish from all the spam out there.

I really like their goal:

"Supporting the brand promises with the knowledge of 150,000 * Know all that we know * Deliver and experience that inspires you * Make a difference * Blow you away with the latest and greatest * Never leave you hanging."

As I reported back in May, Alvenda helped 1-800-Flowers deploy "shoplets" across hundreds of websites for Mother's Day. Shoplets are commerce-enabled advertisements that let customers purchase items without leaving their current website. Jim McCann, CEO of 1-800-FLOWERS.COM stated,

“Through our collaboration with Alvenda, we have now opened millions of stores on Web sites frequented by our customers.”

The results from the program are in, and they are impressive. Using an A/B testing approach, 1-800-Flowers compared shoplets against traditional banner advertising and found that interaction rates with shoplets were 10x higher than standard display ads. Additionally shoplets drove more sales than banner ads, generating 41% more sales per impression. The full report can be found here.

I guess shoppers really liked the convenience of the shoplet.

But the interesting discovery is that interaction rates dropped by 20% when promotional language was included in the shoplet. The less the shoplet seemed like advertising the more customers were attracted. I can think of several ways to explain this, but they would all be guesses. Anybody have a good explanation?

A friend just turned me on to the latest iPhone app to garner lots of buzz. Its called Bump. When two people meet and both have the Bump application running, they can "bump" iPhones to exchange contact information. You can see it work in the YouTube video embedded below.

Sounds like near field communications (NFC), but its not. The bump signals the app to send the GPS coordinates to a server over the internet. If two iPhones send data at the same time with close coordinates, a "bump" is assumed. Then iPhone A sends his contact data to the server, and the server forwards it to iPhone B. And B to A as well.

Now let's transfer the concept to retail. Instead of contact information, let's pass coupons. And instead of bumping two iPhones together, let's just walk into a store and shake our iPhone. That signals the application to send GPS coordinates to the server in the sky, and based on the location, coupons for that store are downloaded. Local, on-demand coupons. Awesome.

But wait, there's more. When you get to the register, "bump" the register to transfer your coupons over. Send your credit card information too, because the communications are encrypted. We just need an internet connection, so no special NFC technology is required.

I might be on to something here...

If you're looking at this article from the blog's homepage, then you can see a bunch of retail articles on the right. They are assembled from all over the web using Yahoo Pipes, a cool technology that combines sources of data from the web. I created a reusable pipe, called RetailFeed, that pulls articles from my favorites news sites and blogs.

What if your customers could do something similar for fashion? The creators of Yahoo Pipes have created a website, called Polyvore, where people can combine fashion pieces into outfits they call "sets." Last month, 825,000 unique visitors either created sets or viewed them. Each article within the set can be clicked, and the user is taken to the originating site, presumably to purchase the item.

The interesting thing about fashion is that no item really stands alone -- they are always combined into outfits So instead of selling individual articles of clothing and accessories as most e-commerce sites do, combine items into outfits. And better yet, let customers do the creative work and then decide which outfits are most popular. This is another way to harness the wisdom of crowds.

As the NY Times reported in their article, there are several avenues for Polyvore to make a profit, but my favorite is to license the technology to fashion retailers so that their customers can create their own sets.

I don't know squat about fashion, so I picked a set from the celebrity list. The Taylor Swift inspired set designed by Karine Minzon Wilson from Brazil is shown to the right. It includes items from marieclaire.com, elle.com, sandalworld.com, and fantasyjewelrybox.com. 774 people liked this set, and it was viewed by about 3,000 people since its creation two weeks ago.

This is a great way to socialize about outfits, and drive additional shopping traffic.

I love the rare bumper sticker that actually makes me laugh out loud, and t-shirts are in the same category. Threadless has some of the funniest t-shirts I've seen, especially their line of Twitter t-shirts. The company has been around since 2000 and has proven their unique business model works. Every Monday they display t-shirt designs submitted by the public. Voters then determine the best designs to print, and the designer gets paid a flat-fee. They offer an online catalog of t-shirts in various categories, and there's always something new. In 2007 they even opened a traditional store in Chicago.

Revolution let's you design you own snowboard, Nike let's you design your own shoe, and Boardpusher let's you design a skateboard, These types of retailers not only offer unique products, but they can learn from their customers and quickly detect trends.

Crowd-sourcing and Mass customization are two great ways in which to better involve customers in retail.

PS- Do you take your phone everywhere?