Case Study: Open Source to Enterprise

When I started at Furniture Row over 11 years ago things were very different. Even though it was after the dot com collapse, the digital age was still in it’s infancy. Google had been scrapping the web for years but wasn’t the make or break tool it is now for digital marketers, iPhones and other pocket internet devices didn’t exist, retail shoppers hadn’t quite picked up on the online pre-shop we all take for granted today, and most retailers minus the huge ‘big box’ stores who were leading the way (like Toys R Us and Best Buy for example) were just getting their feet wet.

Furniture Row, even with its 300+ stores across the US was no different. I came into a team which was also in its infancy as the company as a whole was still a brick and mortar environment. Window clings, building banners, catalogs, and of course giant eight fold Sunday newspaper inserts were still the entirety the marketing strategy, just as it was for most retailers before then and for many years to follow.

One of the first major projects I worked on and led to my rise in the company and to my eventual role as the Digital Creative Director, was the design overhaul of the companies retail sites. The success of the initial redesigns later led to a multitude of overhauls and re-skins to the open source php site that I inherited, and the move to an enterprise eCommerce platform which changed everything.

We knew what we needed to accomplish was going to be much easier said than done. We had a very small budget, a site which had been through three different groups of developers, including one in house and two sets of  outside consultants. The current team was very talented, but as anyone who knows this space, also knows what it takes to rework back-end code structures to align to a new front end design and navigation scheme.

The wire framing tools of today did not exist, and the ones that did left a lot to be desired, so the majority of the UI rework was completed in Photoshop, screen, to screen to screen. I put together around 18-20 comps matching up to each of the page templates we would have to update. There were multiple rounds of back and forth with the developers as Photoshop mockups do not always translate to HTML and browser capabilities. Being over tens years ago, browsers were still the thorn in every web designer and developers side…IE6.

The development team and I worked painstakingly for about nine months from the time of approved comps to the production ready site we released to the public. We learned a lot in the process about how we viewed customers online behaviors, which even as the only user advocates in the company, we were also surprised how users used the site from the page drill down paths, site entry points, time on site, and all the metrics which were available in that very early version of Google Analytics.

As we learned more, more and more users were on the sites, our team grew, executive buy-in began to improve, we were then given more leeway, trust, and eventually a larger budget. The harder we worked to keep up with new user focused techniques in order to create better user experiences, which drove quicker access to products and information and in turn converted to sales in-store and online, we knew it was time to go for the BIG ASK. We needed a very substantial budget and even more trust in the process and us.

At this point we had also started to work an entire new UI design, somewhat under the radar, as we prepped for our next move and it would be the largest undertaking our digital team had been through yet. We were working on building things into our sites that every online shopper and retailer alike takes for granted today, faceted navigation, proper and accurate search, a true price list engine to manage to regional pricing, ability to set up SKUs, kit codes (SKUs made up of multiple SKUs) and more.

Once we had the official buy in to move forward which we accounted to our preparedness while presenting our case, and of course leaning on the trust and growth we had already earned.

Kicking off the process we knew what we needed, but also knew we didn’t know what we didn’t know. We decided that we HAD to allocate a portion of the budget to hire a team more familiar with enterprise level vendor selection than we did.

We chose a company and proceeded on what would become two months of work to set up a RFP that in the end was at least 70+ pages. From there were sent out the RFP and awaited sales teams to come pitch their wares. Each team that came had similar but very different offerings. Each brought at least one solutions engineer to run demos, break down the systems, explain how their code base worked. All were java based but all employed custom layers of their own languages, calls and droplets, on top of the java. Moving to Java made us very nervous not being java developers and had been working in PHP based systems for years.

Believe it or not, and some of you will,  even after having the RFP in their hands, having time to prep demos and explanations for our meeting, one of the companies, which I will not name, was so unprepared we sent them away a few hours into what was to be a full days meeting.

Once we made a decision, negotiated the million dollar plus start up costs and annual license fees, which is standard in the space, we moved ahead. We were convinced that the companies in house team would be a great option and was going to save money too. We found out very quickly that was a bad call and later found out three of the developers where outside consultants anyhow.

After things began to go side ways and deadlines began to be missed we could tell most if not all of the issues were on the platforms team side and not the consultants so we began discussions with them directly, and as it turns out they were all from the same firm so we moved forward with them instead.

The reset with the new team we had to revisit what we had thought was a well planned timeline and move forward using a project manager from the firm. The new plan would take take 9 months, but by the midway point it was clear it would be at least 12. We pushed ahead anyways and went live in about 11. We had many of the same hurdles related to taking Photoshop comps and translating them to functioning HTML and CSS that worked well with our experience goals and back end. Some of our functionalities required customization beyond the platforms out of the box capabilities which was what added to initial timeline being extended.

Over the next few years we continued to improved the sites a few elements at time. We learned even more about customer behavior as we watched closely how the site was being used and started to push harder for online sales, which was part of our original goals but was not a priority. Shipping costs for large goods like furniture was prohibitive, and furniture being very textile in nature, most buyers wanted to touch, feel, and experience the product before buying. The site was having some sales but it was still more of a pre-shopping tool, bringing in better educated customers that gave the sales floor staff a clear idea of what the customer liked.

As that time passed browsers continued to improve, UI and UX were more prevalent skills and techniques and Responsive Web Design (RWD) was the next big thing. We had begun to employ RWD on our web properties that were not part of the eCommerce systems and began to test what was possible. Once we were comfortable we decided to take one of our smaller sites that we had and was also eCommerce as our first entry into combining both worlds.

The major bonus by the time is we started working on the UI/UX with in wire-framing applications available like Balsamiq, the apps had finally stepped up to deliver the help we really needed. Beyond that, jQuery, and prototyping tools like Bootstrap and Foundation where also in their second and third versions and becoming tools we and many others were now using for every project.  Wire-framing in Photoshop was now an old technique, wire-frames were more useful than ever, but leaning on functional low fidelity prototypes was now what we worked on and began using from day one of a new project.

At the time we had a lot of push back from everyone involved. People at the eCommerce platform pushed us hard to ‘just make an app’ using their available tools. Other partners like Scene7 and Bazzarvoice told us it wouldn’t be possible with their systems but I pushed ahead anyways. I convinced both Scene7 and Bazzarvoice that if I could set up what was needed for the front end and prove it would work they would work with us to make it happen. It was a lot of work but I was able to prove to them it would work and a few months later we were the first to live RWD on the ATG Oracle Commerce platform, first to live to Scene7 and Bazzarvoice alike. Shortly after this all three companies were using our sites in their demos and pitches for what was possible.

Added functionalities and processes, like adding local store delivery, free in-store pickup, and better access to the products translated to large increases in over all sales, but even more so to direct online sales. It took time, but nearly 10 years after inheriting the Furniture Row digital properties, making regular improvements within their open source environments, later taking them all to run on an enterprise level commerce platform, adding and building in all the best available user facing technologies, we started to see drastic online sales increases. The digital team was now bringing in sales that rivaled single brick and mortar locations and even bringing in more sales than some of the lower performing ones.