Lessons from changing ecommerce platforms
My company, overstockArt, recently changed ecommerce platforms. It was a complete change of our website, not just look and navigation, but the entire underlying technology.
We switched from the Yahoo store platform (now called Luminate) to OpenCart. The Yahoo store platform was restrictive. We had to build many outside tools to accomplish what we wanted. Some simple tools we were unable to create, such as large images for framed art. Very few developers work with the Yahoo store platform.
We chose OpenCart because it gave full control and full flexibility. Many developers work on this platform. The cost was low and performance potential was high. Moreover, the Yahoo store solution was a service that we did not own. The OpenCart platform is completely customizable and resides on our servers.
As technologies change and as consumers’ expectations evolve, ecommerce companies must constantly improve their websites. Those websites, after all, are responsible for most of the overall customer experience and interactions with our brands.
overstockArt.com, like many other retailers’ websites, has been through many improvements and upgrades. We have added many features over the years. However, in 2013 we decided we needed a major overhaul — a completely new platform. This was a major undertaking.
In my experience, changing ecommerce platforms can end up in one of three ways.
- Disaster: Large amounts of money and time spent and you can’t even use your website. Sales may even decline due to lack of focus.
- Minor impact: You launch a new website and nothing changes — sales don’t increase and shoppers don’t responded one way or the other.
- Major improvement to sales and profits: The new website increases customer engagement and conversions, providing more sales with no increase in marketing and other expenses.
All merchants presumably want to be as close as possible to point 3: growth in sales and profits. But re-platforming projects often fall somewhere in between points 1 and 2.
When considering new platforms, ecommerce entrepreneurs typically focus on option 3: the potential for major increases in sales and profits. This is also what technology vendors will focus on.
The questions I want to address in this post are:
- How to guarantee or at least increase the chances of ending up with option 3?
- How to prevent disaster, as in option 1?
In fact, we tried re-platforming overstockArt many years ago. It ended up in disaster. We paid a lot of money and got nothing for it. We ended up in court and lost.
Our experience this time was far from smooth. It took longer than expected, with many flaws. Some of my recommendations in this post will be what we, overstockArt, should have done.
There are no guarantees that your shoppers will love your new website and that the adoption by those shoppers will be smooth. However, based on my experience, it is guaranteed that during this long project you’ll have problems with your web development vendor. And not all of your desired features will work smoothly. Finally, it is guaranteed that it will take longer than expected and cost more! And to make it worse, your friendly web-development vendor will likely blame you for the increase in time and money.
Our current development vendor has done a decent job overall. The original vendor (not the current vendor) told us initially that this project would take 90 days, which we knew was not possible. We assumed it would take one year — four times the vendor’s original estimate. It actually took two years — eight times the vendor’s estimate and double our estimate. Finally, the staff of the original vendor was unable to finish the project. We took it from them and hired other technology firms to finish the site.
There are many ways to look at a project of this size. Certainly we made many decisions along the way. Here is what we’ve learned.
- Why should an online retailer consider re-platforming? There are three simple reasons for this or any other major investment: (a) Will it increase sales? (b) Will it decrease operational expense? (c) Will it reduce inventory? If the answer is not yes to at least one of these, there is no reason to consider the project.
- Assume the worst. Prior to moving forward, a retailer should know that it will cost more and take longer than estimated. Many web development firms base their profitability on change orders.
- Complexity creates errors. The more complex your website is, the greater potential for error. We took all of the features and developments we created over 10 years and added them to our new website. This created a lot of complexity to the project. Even though our vendor knew of those features, the complexity and the amount of features made the project very difficult.
- The planning phase is most important. You know your business better than anyone. Make sure to educate your web development vendor so that the architecture of the new site suits your business now and down the road.
- Maintain close communication with your vendor. Once the project is launched, insist on regular updates and dialog — negative or positive. Be frank and transparent with your vendor.
- Involve your entire team to some degree. Have your staff test the website’s frontend and backend — reporting bugs and missing features. Do this throughout the process.
- Cut your losses. If the project starts lagging, and you feel your vendor is becoming unresponsive, cut your losses and take over the project. Publish the site in a beta environment, push it to your servers, and use Git (the open-source software control system) to involve multiple development companies to complete the final part of the job.
- Test on multiple browsers. Test your new site not just with the newest browsers, but also older versions of the major browsers. Also, test with multiple devices — smartphones, tablets, laptops, desktops.
- Test key areas and functionality.
- Place an order with multiple modes of payment.
- Can customers use coupons?
- Does product-page functionality work?
- Are the search results relevant? Conduct many searches and see if the results make sense.
- Ask friends to browse the website and provide feedback.
- Test site speed for both desktop and mobile. Speed is a key aspect of conversions.
- Transfer all the data possible from your previous store.
- Product information and images.
- Customer information, including accounts, order history, reviews, and wish lists.
- SEO data: Meta information and 301 redirects.
It’s also a good idea to place a banner on your current website and ask customers to browse the new, beta website to provide feedback. Customers appreciate being included. By asking them for their opinions, you’re not only getting helpful information, you’re also building stronger relationships.
Finally, remember re-platforming or a major website upgrade is a long-term play. Your sales may actually drop a little as shoppers are getting used to the new website. But the rewards can be amazing.
Here are our results after launching the new overstockArt.com site.
- Month 1: Disappointing sales. We experienced a year-over-year sales increase of 15 percent. But we expected 30 percent based on previous months.
- Month 2: Major improvement to sales, but still not at target. We had a decrease from the prior year.
- Month 3: Sales increase still not as projected; growth was single digits. By now our website was bug free and started looking like a major leap from our previous site.
- Month 4: Conversion rate increased by 60 percent. Sales are up even more than that as traffic increased as well. Our website looks amazing. Finally, our dollar-per-visit matrix is up significantly, allowing us to advertise more aggressively.
In short, a complete re-platforming carries much risk. But the results can be terrific.
Have you changed ecommerce platforms recently? Are you considering changing? Please share your thoughts.