Offline and online brands were initially kept distinct, then they were awkwardly merged. Initial e-commerce efforts were flashy brochure sites, with rudimentary shopping carts and checkout
Then systems that tried to anticipate needs and accelerate checkout were deployed.
From Google's perspective, though, one of the biggest changes has been the growing importance of search, both at the site and the Web levels.
Google officials point to a not surprisingly huge number of purchases that follow Web searches as well as abandoned shopping carts that immediately followed a non-productive site search. A classic example: A visitor searches a retail site for "video camera" or "movie camera" and finds nothing and leaves, when the site was expecting "camcorder" and would have shown him 20 models had he used the magic word.
Two Google executives this week talked about their view of e-commerce's future and how IT executives should plan.
Seeing tons of additional details from retailers posted online will be one of the first changes expected by John McAteer, head of Google Retail. ("Yes, 'head' is my actual title. Better say 'head of,'" he said.)
McAteer's view is that a lot of merchants will pour continuous structured feeds of data?including SKU listings, daily inventory and hours of operation?into public search engines such as Google. This is currently being worked with on Google Base.
In theory, this would allow much more specific and relevant search results and?also in theory?much greater e-commerce revenue for the retailers who cooperate.
Not only could a consumer seeking a particular model of electric drill see the retailers who claim to sell it, but also the closest merchants that are open right now and that apparently have five still in stock.
How often does McAteer predict those inventories will be updated? "The more they update it as often as possible, the better off they're going to be," he said. "Realistically, we can peruse that daily. For some retailers, it will be weekly."
The shop-online-and-buy-in-stores efforts will continue, McAteer predicts, but having much more specific information will make such efforts perform much more smoothly.
Some retailers?such as Best Buy?have tried to insert humans in the middle of the process to make the system more accurate.
That approach calls for customers to use the Web site to make the purchase and identify a local store that apparently has the item in stock. The site then tells the customer to stand by for a few minutes and wait for a confirmation e-mail. A beeper then sounds in the store and a store employee runs out onto the floor, finds the item and gives it to someone in customer service. Only then does the customer receive the e-mail confirming that the product is at the store.
That approach, McAteer said, simply couldn't scale sufficiently and will likely fade away. That will put the pressure right back where it should be, on technology systems to more accurately know what is and isn't in stock. That might require inventory data to be fed to search engines?for appropriate referrals?along with possibly POS (point-of-sale) and RFID item-location information.
The POS data might be critical during high-volume purchasing periods such as the end-of-year holidays. If a store, for example, has only one or two of a much-in-demand product, it might be helpful for the site to know that one was just purchased 5 minutes ago. The RFID data could indicate if the item has left the shelf, either already sitting in someone's cart or perhaps simply misplaced.
By using search engines as makeshift "inventory visibility tools," McAteer argues, retailers could find search engine referrals sharply increasing their revenue by allowing local finely tuned searches.
That said, the practical issues that retailers face today pose some huge obstacles to this kind of data sharing, McAteer said.
Fears about sharing proprietary information in a way that direct rivals could easily access is certainly going to be a psychological barrier, he said, but it's one that will likely fade as more of the industry makes the move.
The potentially biggest hurdle is that many retailers today simply cannot provide that data for Google because they can't easily access it even for themselves.
"Google is one of the best companies at organizing the world's information," but it can't organize what retailers can't provide, he said.
"Retailers can't even give us what is the stores today," McAteer said. "I have had dozens of meetings with C-level executives at retailers," and online data control "is only one portion of what they do everyday" and would not likely be a near-term priority.
"Some will make it a priority," he said, but "I think they see it as something not as important to them as it should be."
McAteer said most retailers see search as a good distribution channel, but they have a hard time making it a strategic priority.
"Is there an ROI for it? Do I have the resources to do it internally?" he asked. "They either don't see the opportunity or they don't have the resources or both."
What will ultimately motivate them? "Their competitors will do this."
The corporate priority is not merely to make the data available, but to have it cleaned up internally. Many retailers, McAteer said, couldn't do this today even if they wanted to.
"They can see at the corporate level when Crest moved off the shelf in Palo Alto, but they can't push the data out," he said.
Google's plans also call for Web analytics for retailers and helping to identify an accurate list of synonyms and likely spelling errors for their internal searches.
But McAteer isn't the only e-commerce future-predicting executive engaged in Google Gestalting.
In another part of the Google galaxy sits Deep Nishar, a Google director of product management who is focusing on the handheld/smart phone/mobile environment.
As PDAs and cell phones become more sophisticated in the ways of Web and SMS, they are expected to start having a significant impact on e-commerce.
That isn't a reference to literally using the PDA/phone as a payment device?although that is likely to happen as well.
No, it's a reference to what many in e-commerce believe will be a very significant change for e-commerce in the next year or two. It's what will happen when the online world starts regularly intruding into the offline world.
Envision a consumer pushing an oversized cart down the aisles of Costco, while searching for items on Costco.com. No need to try to find a rep to see if the store has a particular microwave size and color in stock or to see a demo of the product.
"Is this really a good price?" the hypothetical consumer asks, seconds before clicking to one of several price-comparison Web sites. The consumer's PDA/cell phone will likely have a bar code/RFID reader to make such comparisons faster and more precise.
Or that consumer may be inside a crowded mega-mall during the height of holiday shopping and will use her mobile device to help determine which store to shop in based on an almost real-time inventory report.
Although such a near-future view has been discussed in the offices of most of the major retailers, very few have been actively programming and designing Web pages with the mobile user in mind.
When will mobile e-commerce be the norm, with a healthy percentage of purchases being done on a mobile device? That's hard to say, said Google's Nishar, and it will be very different in parts of Asia (which are furthest ahead in the mobile acceptance scale), Europe (middle) and the United States (which is far behind).
In the United States, "it could be in two years, it could be 10 years," Nishar said.
But that's for it being the norm, with mobile purchase experimentation already showing up.
"People are already starting to do transactions" on mobile devices, Google's Nishar said, so it's critical that retailers start "making sure that they are not just building for the PC Web, but they focus on the mobile Web, too."
Those changes include GUI issues (format assumptions that can include much smaller screens), bandwidth restrictions (less information overall, but a heavy reliance on text for what is needed) and, most critically, the selection of information to be displayed.
That last element requires companies to think through the kinds of datapoints that will be helpful to people when they are using mobile devices.
"When a user is on the go, what is the information they care about? They are most likely not going to want images from three different directions," Nishar said. "They'll want very quick and easy checkout, autofill and a whole host of other things."
That mobile hybrid, according to Nishar, will likely be a telephone, so allowing a user to conduct some of the transaction automatically on the phone is sometimes the most efficient way. A computer?especially one with the keyboard and screen as small as today's typical mobile devices?may not always be that efficient, he said.
Nishar argues that consumers and businesspeople are not likely to ever think of PDAs (or PDA/phone hybrids) the same way they think of laptops or desktops.
"I don't see them changing that behavior any time soon. Realistically, when people get their phones in their pockets, they don't think of it as a computer," he said. "In the Asian market, where many people don't have computers, they still just view that phone as a mechanism that is valuable and can get them information quickly."
Nishar draws a psychological distinction between browsing and looking, where "looking" is a more finite attempt to get a specific answer to a particular question or problem. It's more informational than recreational.
"They might use it to figure out the latest crop prices. That's not like browsing for movie reviews," he said.
This isn't to say that entertainment information isn't important, as Nishar is quick to add that, "in general, entertainment is by far one of the top categories of what people look for on their phones. But people are not browsing on their phones. They are up and about and they need to find a local restaurant."
Again, he points to a sharp difference in how mobile content will be formed in Asia versus Europe and versus the United States, something global retailers must understand.
"We have to look at the context. In the Western world, the consumer is much likely a lot closer to a bigger machine," he said. "If you look at a country like Japan, anywhere from 20 to 25 percent of all Internet usage is now happening on mobile phones."
He cites the mobile Web site of United Airlines as a good example of where retailers should be headed.
"It's very functional," Nishar said.
One of the most ballyhooed applications for mobile commerce?as a opposed to mobile e-commerce?has been to use the phone's location-awareness to drive sales, such as telling the consumer that there's a sale going on at a merchant 200 feet from where he is standing and the sale will end in 30 minutes.
With rare exceptions, those applications have simply not happened in the United States. There are two reasons for that, Nishar said, even though the capabilities have been touted since the late 1990s.
Initially, the delay was from the cell phone carriers who "hadn't figured it out," he said, given the complicated technology issues.
"But today, they have figured it out. The technology has been there for the last two or three years," Nishar said, adding that the second delay has been "a whole host of issues involving privacy."
The fear among carriers was that consumers would be frightened of technology that would take their exact locations and share it with various merchants and their employees.