To everything there is a season, and a time to every purpose under Heaven.
There is a time to sow, and a time to reap…
This Biblical verse is true in agriculture, personal life, as well as merchandising and marketing. You can have the best new ‘feature’ to your web site, but the timing of the release should be carefully considered. If your feature is of minimal impact to the navigation and overall familiarity (look and feel) of the site, then there is little risk in implementation. If, however, you are planning a complete redesign of the presentation and business logic layers, you may consider the time of year for the release.
Why: One retail chain that I had worked for decided that their website needed an upgrade. That was an understatement - the existing site was very dated in both style and functionality. Many features that the underlying web framework offered were inaccessible due to a prior rushed upgrade that ‘broke’ what is a very robust J2EE-based framework. Business and IT knew that a complete rewrite was the only way to efficiently keep the site extensible, maintainable, and scalable. The initial launch date was early September, with a fudge factor of a couple of weeks.
As the months of planning a preparation, design, coding, and testing went by, it was clear that the target date would be missed – by about 2 months! That brought the time of the upgrade to the apex of annual holiday sales season. While this may initially sound like the perfect time to release the site to the customer base, a little more thought might suggest that the opposite is true – it may be the exact wrong time.
Who: Humans are creatures of habit. We like familiarity and loathe what is unfamiliar and different. We often come to embrace the new and different over time, but our initial reaction to change is typically resistance. Fortunately, the web offers most of us scores of choices when we are searching for information or products to purchase. There are always options.
If my favourite tech site (http://www.tigerdirect.com) changed the complete navigation, design, and business logic, I might consider purchasing from my second favourite site (http://www.newegg.com). While I do not like the look and feel of New Egg (a personal preference, not a critique on their site design), I have grown use to it and can find what I need fairly quickly. Tiger Direct will lose my business until I have the time and need to re-learn how to find products.
NOTE: Tiger Direct has not changed their site; it is easy to navigate, has great pricing, and their product delivery is stellar.
What: Data translation is very tricky, and more often than not, it results in some type of corruption on the first attempt. There are many ways to insure that data will be intact after the conversion, but none are 100% reliable. The rule: always plan for a data layer error during and soon after an upgrade.
When: When, then, is the optimum time to do a complete site upgrade? During the least busy time of the year. For e-commerce sites, that time is often between mid January and the end of April. Plenty of time to work out the kinks in the new system.
Where: Do you have the resources and skills to perform the upgrade internally? Is there adequate personnel in house to both maintain the existing site and code the new site? Do you have adequate testing personnel, with the proper testing tools, to insure that what is being done will really work? Is outsourcing your best option?
These questions should be answered objectively and without regard to pride, personal advance, or budget pressures. A team that is not sufficiently trained or does not have the time to code on such a scale will experience mistakes, delays, stress, and, ultimately, failure. Augmenting an in-house team with consultants who are experts in e-commerce is another option. This works only if your personnel are eager to, and have the capacity, to learn the details of an e-commerce framework. It is one thing to maintain and enhance a site, and a completely different process to code a site from the ground up.
How: Do you flip a switch when performing a complete site upgrade, or do you gradually test the site with a subset of your visitors? Obviously, you direct a percentage of your customers to the new site to assist in production testing. You can ask customers to ‘opt-in’ by presenting them with a link or popup when they first arrive at the site, or, redirect nn% of the traffic to the new site automatically. The opt-in approach gives the customer control, which often results in a positive reaction.
Along with the opt-in, provide the customer with the option to revert to the ‘old’ site if they cannot find what they are trying to purchase.
Once the guest has visited the new site, ask them to fill out a 3 – 5 question feedback form. One of the five questions will be a text box so they can give you as much feedback as possible. Again, if they are involved with the process, they will more likely embrace the new site, over time.
As a matter of safety, the cart/checkout process might remain as-is initially. In other words, your marketing and merchandising of products can be upgraded, but once products are purchased and the customer is checking out, the familiar cart gives them a sense of security. Keeping this critical component the same during testing reduces the likelihood of abandonment and order corruption.
Conclusion:
Given the strained economic climate, there is an ‘out’ if the new site fails. The economy can be blamed over the timing of the upgrade release, and no one will be the wiser – unless the site is completely down, you look at the analytic data and see that the abandon rate increased during this period of time, that orders were not committed to the database, or that products were out of stock and still presented, etc.
So…
- Timing is (almost) everything.
- Be objective – new is not always better. Emotion, self aggrandizement, and pride should never drive a business decision of such magnitude.
- Research your customer base to determine what they will accept (and when).
- Look at your competition and research their successes and failures concerning feature enhancements and upgrades.
- Dates are numbers that exist on calendars – they can be changed.
- Skills can be learned, but complex skills often cannot be learned quickly with great aptitude.
vr