E-commerce is much celebrated as a successful application of technology for present and future. But even in technology and Internet timescales (and certainly in business timescales) it is very much at the beginning of its life. Mistakes are made (and not always admitted to) and lessons are still being learned.
Here, some of the big names in New Zealand's small e-commerce pond think back on the mistakes they have made and the times they have had to change direction. Clearly, some are hesitant to admit to past errors, but willing to dish out good advice, nevertheless.
Flying Pig boss Mark Battles says, on reflection, the e-tailer should have gone with a more modular approach in developing its systems.
"We took the Whitcoulls site over and put it under the eye of a technology partner in Australia, then gave it to Advantage."
Advantage developed a new site on the basis of that, notably to include a broader range of products.
The original development, however, was from scratch, so it could be said that Flying Pig inherited that development's thinking.
"By the time you write a technical specification [for an e-commerce site built from scratch] it's already out of date," Battles says.
"The business requirements change, and that drives a technology change." Added to that, new technology is coming through all the time.
So Flying Pig should have gone with an out-of-the-box solution and modified it to improve the fit between its capabilities and the company's needs, he says.
Then the company should have acquired best-of-breed applications for specific tasks like catalogue management, content management, and the management of an affiliate programme (affiliates are external companies' Web sites that drive traffic to the Flying Pig site).
"A best-of-breed applications can help you establish that relationship and then count the clicks coming in from affiliate sites."
By comparison, an original application in that area is only a partial solution.
"We decided to go for a bespoke solution because we had confidence in the ability of [the] Advantage [Group].
"But we were the first really high-volume site it had done, and we were in some ways being used for research."
Another key lesson is the importance of automating the back end of the operation and integrating it adequately with the Web front-end, he says.
In the initial stages of Flying Pig's exposure to the public "we had a lot of manual processes to manage order procurement and fulfilment", Battles says.
"Exo-net has now supplied us with Web-enabled ERP. Phase one was just the financials, and a few months ago, we took on phase two, to automate the Web ordering.
"On reflection we should have accelerated that integration; we should have invested more in it."
The reason for not doing that, he says pressure for an early launch, a factor that has clearly affected others.
Thirdly, it pays to stay away from the bleeding edge, Battles says.
"We were one of the first sites to use Windows 2000, and one of the first to try out Microsoft's hardware clustering. We were heavily involved in educating the market."
Flying Pig should have stayed away from that edge and "opted for the more tried and tested" solutions, he says.
Portal operator eforce's major mistakes were on the business side rather than with technology, says the head of the company's portal division, Steve Pike.
The lesson for eforce was to do more market research and establish what you're really good at.
"Originally, late last year, we envisaged eforce as building an online community, and then using that critical mass to exercise bulk purchasing power.
"We built the database very quickly, but when it came to supplying [the people in the databases'] needs, we realised we had rapidly moved into the business of e-tailing.
"We had to develop sites [for selling various kinds of goods and services] and we had to source products."
The company's first development was Power Choice, giving consumers the ability to switch electricity suppliers online. This remains one of the portal's offerings.
Then there was Co-buy, which approached the bulk purchasing exercise head on.
"The challenge there was that the population of users who buy on the Internet in New Zealand is low," so major bulk-buying power could not be exercised to get very low prices from vendors.
"We didn't have the expertise in retailing, much less in e-tailing."
Co-buy is on the way out, he says.
"It will continue to run for two or three weeks," he says. "Then we'll replace it with something else."
In the process of trying to sort out its supply lines, eforce acquired Product Sourcing International, a specialist in supply-chain management, as well as Web design specialist Imagic.
Now there is an opportunity to look at the total group and exploit its capabilities.
"We will leave e-tailing to the specialists and focus on the value of the database [of potential buyers], targeting them with online promotions for goods in which those specific buyers have expressed an interest."
PSI will also concentrate on sourcing goods for bricks-and-mortar retailers. It will use the Web too, but this will be a separate exercise.
"The principle of 'build it and they will come' is not necessarily true. Internet users are sophisticated and educated people. They know what they're looking for and the market is unforgiving.
If you can't supply their needs, then they can quickly go somewhere else, Pike says.
"You have to know your market; you can't just flog goods indiscriminately to the masses.
"Our biggest mistake was in trying to compete in the retail market, when we didn't have the expertise; and furthermore, trying to e-tail it.
"We were trying to source products ourselves. That's something better done by a competent partner."
More generally, Pike says, the lessons are: "Do more research to establish what the market you intend going into is like.
"You have to know, or realise quickly, what your strengths are.
"Our strengths are in building an online community and matching buyers and sellers.
"The New Zealand market is too small to e-tail one product or one brand. Well, maybe some people succeed in doing that," he reflects, "but not us."
The "first to market" drive is a strong one, but an e-commerce organisation should resist it, and spend more time considering what it's good at.
"And be quick to react to what the market is telling you."
Eforce spent a lot on marketing, giving away iMacs and the occasional Volkswagen.
"In hindsight, I don't believe we needed to give away that much," says Pike.
There were warning signs, in the shape of several companies in the US that had gone broke through excessive marketing.
On the technical side, the development was a happier and more successful story. Eforce concentrated on Microsoft Active Server Pages as its technology of choice.
"We were attracted to ASP for its rapid development potential and for the skills that were in the market, which would make it much easier to support," says technical specialist Mike Dempson.
This approach has paid off.
"We've pretty much been lucky in our choice of platform." That is a powerful IBM processor with SQL Server.
Be careful not to underestimate your development time, says Bronwyn Evans, managing director of beauty products site Beauty Direct.
"We have found a lot of development houses are not good at doing proper functional specifications; they underestimate the difficulty [of the development task].
"We made business plans around their promised schedules and then they didn't happen; so things got delayed."
Skills in design of a series of Web pages do not necessarily go with technical skills which allow them to estimate the difficulties involved in bringing that vision to fruition.
"There is still a lack of focus on the needs of the customer by people who design. A site must be convenient for the customer to use; many designers are focused rather on how cool it looks."
Beauty Direct's staff ended up doing some of the work that the outside developers should have been doing, she says. And it wasn't just that the company picked an isolated bad developer.
"I've seen quite a few and they were all like that."
Evans herself has had experience of Web site design and development at Telecom. If she hadn't, she says, things would have gone more badly wrong.
The lesson from those experiences is that "you have to take control of the project", and not leave too much to third-parties, Evans says.
You should evolve a good business model and understand it.
"You can't just put things out there and hope people buy them; you have to know what your selling proposition is. Ours, we decided, is convenience; if people can't get to a store [that stocks the product they want] they don't want to wait a week for delivery."
Getting a quick and efficient site that handles customer needs is a matter of site design, and that is a learning process.
Evans admits "we haven't got our site right yet".
It is attracting the kind of customer who knows what they want and can find their way to it.
"But we want to be able to lead the customer [who lacks that knowledge]."
There have been almost no mistakes in a technical sense, she maintains.
One false trail, however, was to put the database of products in the front end of the system, so it was actually "on the Internet" and at risk of interference from outside.
Now the database is off-line, and when a new product needs to be added, staff "dial up" the database separately, and transfer the data through ODBC.
Another benefit of the structure is that customers' credit card numbers are stored off-line and do not have to be encrypted.
"Some customers were complaining that they needed to re-enter their credit card number every time."
Now they will not have to do that.
Businesses still show a low general understanding of what the Internet can achieve for them, says Evans.
"And I think they're scared of losing control."
They have controlled circulation of their mail-order catalogues, so they are hesitant to expose a catalogue to everyone.
But control can still be exercised by the look of a site, and appropriate pricing, to attract the kind of customers you want, she says.
One of the vital challenges is to steer a middle course between the hype about the different world of the Internet and sound corporate practice, says Bruce Gordon of E-Loan, the first service founded by the eVentures group.
"Never forget it is a business," says Gordon.
"It must be offering services that someone in the supply chain is willing to pay for."
All the business disciplines, from sound financial practice to sound project control, should still be applied.
"It's not such a new world that these go away."
But at the same time, development for e-commerce is different.
The structure of committees that assess one project after another in detail does not apply.
"Lots of things will be working in parallel.
"There is not the rigid structure of demarcation between roles as you get in a corporate - you draw upon those that have the capacity and experience you need at that time.
"Structure is still important," he acknowledges, "so the hard calls can be made when they need to be.
"Democracy is an important cultural ethic to debate issues quickly and if the team is any good the passion will be intense.
"But in the end a decision is made by the leader and everyone supports that decision going forward."
Furthermore, you will be far more accountable for your mistakes in an e-commerce environment than in a bricks-and-mortar office or shop, he says.
"If a conventional retailer puts something on the shelf that doesn't sell, he can discount it, or just take it to the dump.
"If you make a mistake on the Internet, thousands of people will see it before you do."
Hence quality control is perhaps even more important on the Web than in a conventional business.
"Secure the URLs you want when you have the idea; don't wait until you get going," he advises, having had a lesson in this from the company's US franchiser, which failed to secure eloan.co.nz, so the company has to operate under the name eloannz.co.nz.
"It should have ensured that it was reserved in all relevant territories," he says.
He counsels prospective e-commerce developers to "reuse whatever you can from partners or owners - speed and experience are critical to success".
In contast to eforce, he thinks being first to market matters, since the New Zealand market is small.
That said, you should not think only of your own territory; the Internet is a global marketing opportunity.
And a final point: "Negotiate a discount from the local food suppliers prelaunch, because your team will be working long hours. The cliche‚ of pizzas is actually true."