Case Study: Winning With E-Commerce
A 2001 case study in surviving the dot-com shakeout: how one retailer fixed checkout friction and matched spend to buyer intent instead of chasing traffic.
2001 was a brutal year to be selling anything online. The dot-com collapse that began in spring 2000 was still working its way through the industry, and the venture money that once poured into any site with a shopping cart had mostly dried up. Amazon itself did not post its first profitable quarter until the final months of 2001, after years of losses that had made “get big fast” the default strategy for e-commerce. The businesses still standing by mid-2001 were rarely the ones with the flashiest sites. They were the ones that fixed the boring parts: checkout, trust and unit economics.
Picture a mid-sized specialty retailer (a composite of the kind of business common at the time, not a real named company) that spent 1999 and 2000 building an ambitious catalog site with hundreds of products, a custom shopping cart and a marketing budget aimed mostly at display banners. By early 2001, traffic was healthy but sales were not. Cart abandonment was high, repeat customers were rare and the cost of acquiring a first-time buyer through banner placements had crept past what that buyer was worth.
What changed
The turnaround did not start with a redesign. It started with basic instrumentation: server-log analysis, still the standard way to trace visitor behavior before hosted analytics tools became common, showed exactly where people dropped off. Three problems stood out.
First, checkout forced customers to create an account before they could even see the total with shipping added, a common practice at the time that quietly killed conversions. Removing that requirement and showing shipping costs earlier cut abandonment noticeably.
Second, the site had almost no visible trust signals: no clear returns policy, no security notice near the payment form and no customer service phone number. Adding a plain statement that checkout was encrypted (SSL was becoming standard, but customers still needed to be told it was there) and a clearly stated return policy measurably increased completed orders.
Third, the marketing spend was scattered across broad banner buys aimed at anyone who might notice them. That budget shifted toward the pay-per-click model GoTo.com had pioneered a few years earlier and would soon rebrand as Overture, targeting people actively searching for specific products rather than browsing at random.
The result
None of this was glamorous. Over two quarters, the retailer’s conversion rate roughly doubled and its cost per acquired customer fell by a similar margin, simply from removing friction and redirecting spend toward people who already intended to buy. Revenue growth slowed compared to the free-spending days of 1999, but for the first time the unit economics worked: each new customer cost less to acquire than they were worth.
The lesson for 2001
The dot-com crash punished businesses that had optimized for traffic and funding rounds instead of profit. The e-commerce operations still standing by mid-2001 shared a few habits: they measured the checkout funnel rather than just visits, treated trust signals as a conversion feature rather than decoration and matched spend to intent instead of reach.
None of that required a large budget. It required discipline: fix the leaks in the funnel before pouring more traffic into it, and spend marketing dollars on people already looking for what you sell rather than people who might glance at a banner. That is a smaller, less exciting story than the one the dot-com era told about itself, but it is the story that produced businesses which lasted past 2001.
Keep exploring: browse more Marketing Strategy guides, see everything we published in 2001 or check out the Digital Business Marketing Awards.
