Building an e-commerce business: Lessons on moving fast

6 April 2020

With consumers moving online in reaction to coronavirus restrictions,
companies will need to learn how to launch new e-commerce
businesses quickly.

While the full implications of COVID-19 are

still unknown, it’s clear that the impact on retail is

already significant. Emerging evidence points to

a significant shift, as customers scale back their

shopping in stores and instead go online.

In China, online shopping has increased 15 to 20

percentage points,1 and e-commerce in Italy has

increased 81 percent compared with the last week

of February.2 US consumers have largely followed the same pattern. The COVID-19 crisis is first and foremost a human tragedy, requiring companies to take immediate steps to protect their people.

Amid this human cost, companies are also starting

to come to terms with the impact of the crisis on

their businesses. With offline shopping collapsing,

companies’ strategies will need to focus on

fortifying their web presence and, in some cases,

building an online business.

We have found that companies can create a

working e-commerce site in much less time than

they think. In fact, a poor understanding of what’s

really possible with digital can be one of the most

significant hinderances to moving quickly. As we

highlight in Fast Times: How Digital Winners Set

Direction, Learn, and Adapt (Amazon Publishing,

February 2020), low aspirations lead to modest,

even negligible, outcomes. In our experience,

completely new businesses can be launched in

fewer than four months. Normally, a company

requires at least six to nine months to get a complete solution up and running (and often considerably longer when outsourcing the work to multiple

vendors and agencies, with the increased need for

coordination slowing down cycle times).

We had the opportunity to watch this thesis play

out in the real world recently at one European retail chain that has around 1,000 brick-and-mortar

stores across the world. The chain, owned by a

private-equity fund, had no e-commerce presence.

Although it had previously considered e-commerce, there were serious concerns about whether it could ever work, given the assortment, concept, and even brand constraints. Despite these issues, a variety of pressures, from consumer demand to competitive constraints, forced the company to take action.

Thirteen weeks later, it had a functioning

e-commerce business in one of the regions it

operated in. Not only that, its launch was successful from the first month, generating almost 3 percent revenue growth within the chosen region, tripling average basket size compared with retail stores, and maintaining a high customer-satisfaction score.

When COVID-19 started disrupting daily routines

in Europe, the e-commerce revenues jumped up

threefold almost overnight.

This article describes the main lessons from

that program.


Be pragmatic

Before work started on launching an e-commerce

business in our example, the company’s leadership gathered to discuss the ambition for the endeavor. One word dominated that discussion: “pragmatic.”

Rather than attempting to launch a full-blown digital business across all markets at once, the CEO opted to go to market fast with a limited offering and in limited geographies, gain strong traction, and

then scale up and out aggressively (see sidebar,

“Supporting your remote teams”).

Within days, a new team created a comprehensive, week-by-week plan that covered everything from creating customer-testing touchpoints to setting

up the warehouse with electricity and equipment

to photographing and writing copy for every one

of the 800 products to go on offer. All nonrelevant

initiatives were postponed in favor of efforts that

had direct customer impact. The team closely

scrutinized every feature and ruthlessly prioritized

intermediate release goals for what mattered most.

This biweekly review exercise also made room for fixing problems when things inevitably went wrong.


This lesson was repeated a second time when

management reevaluated scaling plans in light

of COVID-19. The team focused instead on

further strengthening the core functionality of

the e-commerce business and providing an easy,

dependable way for a locked-down population to

shop for what it needed.


Be clear with responsibilities: Assign

ownership, not tasks


A crucial element in enabling speed during our

example retail chain’s launch of its e-commerce

business was clearly designating which teams were responsible for which tasks—and then giving them the space to complete their work. Management

created four teams with responsibility for specific

work streams and ownership over a certain portion of the customer journey. Management then stepped back, giving teams the responsibility and flexibility to solve every unplanned issue that occurred there and pushed them to be creative with solutions. The four teams and their tasks were as follows:

— The tech-and-design team was in charge of defining the microservices architecture, including codesigning the online-store concept with customers and building it iteratively. It also had to develop tech capabilities for connecting the back end of the store with the warehouse management system, inventory synchronization, and order handling. The team owned the customer journey from the moment the customer arrived to when an order landed on the handheld tablets carried by the pickers in the warehouse.Over time, the tech stack developed to a point at which almost all content-level changes could be done without developer involvement.

— The operations team was in charge of setting up the warehouse: establishing packing stations and picking trolleys, setting up the workforce for the warehouse and customer support, liaising with carriers and intermediaries to set up a new relationship, and detailing the procedures for handling all common and edge cases. It owned the part of the journey that started when the warehouse received an order and ended when the customer received their package.

— The product-assortment team analyzed the full product assortment available, chose the 800 best SKUs to launch with (based on multiple quantitative and qualitative criteria—for example, “shippability” of items such as ice cream and fragile glassware, sizing to fit existing boxes, and logical fit with other products in the assortment); obtained samples to photograph, measure, and describe; and created creative online-only bundles. It owned the portion of the customer journey covering all of the steps necessary for a customer to find, understand, and choose a product. As such, the team worked closely with the tech-and-design team in the initial phases.

— The marketing team, which kicked off nine weeks before the launch, created a detailed marketing launch plan, set up a structured customer-relationship-management system and biweekly campaigns to increase the size of the database by almost threefold prior to launch, worked with an agency to come up with a creative launch campaign, and planned an operation in which all the content and merchandise (such as posters, T-shirts, and bags) would roll out in unison on launch day across offline stores, social media, search advertising, influencers, and various PR channels. This team owned the customer decision journey, from complete lack of awareness to the moment a customer landed on the home page for the first time.

At the end of each day, all the teams came together for an all-hands checkout to update each other about the newest developments. This habit—along with solid agile ceremonies, such as weekly sprint planning, biweekly demonstrations, retrospectives, and use of collaboration tools (Kanban boards and Slack)—was the perfect counterweight to workstream independence and ensured that everyone was in sync at all times.

Learn and adapt

Putting in place the right measures and key performance indicators early in the process of creating an e-commerce business is as important as launching quickly. It allows companies to track the progress that matters so that they can learn, adapt, and drive continual improvement. It is not enough to measure overall conversion or conversion by channel. Instead, companies need more granular metrics—for example, to identify relevant customer cohorts, measure microconversions per cohort, and then improve for that use case continually. Metrics or key performance indicators that are too high level don’t provide a clear-enough story. This concept, in particular, is crucial because launching an e-commerce presence is not a discrete project but much more a program of continual improvement.


In our retailer example, the leadership decided early on that the goal of the e-commerce initiative was to launch quickly a business that could gain traction within a single region and then to use it to learn and improve before going all out globally. The company knew that it needed first to build a muscle to operate and optimize a multinational digital business—and that the best way to do that was by learning through doing. After a successful launch, for example, the team spent a month analyzing pain points and bottlenecks and then experimenting with new solutions to improving the experience continually. It was able to act quickly because it had established a simple spreadsheet model to track progress. The team checked it daily the first two weeks and weekly thereafter. It then added new metrics as goals changed. For example, when focusing on growing B2B segments, it tracked the number of orders and revenue for B2B customers daily. When focusing on improving operations, it tracked pick-and-pack speed and the percentage of fulfilled orders per day.

The team also developed prototypes for various new features to test with customers and abandoned many pet ideas that simply didn’t stick. The team had to go back on one of the online-exclusive deas—staff curated bundles of products based on common occasions, such as birthdays— that tested very well with customers prelaunch but simply didn’t perform in reality.

By keeping a weekly sprint rhythm, it was possible to pivot away from poor ideas and generate new ones constantly until the team found ones that worked. Week by week, the improvements and lessons accumulated, and by month three, the e-commerce business had matured into a state at which a good sales day wasn’t a victory but business as usual. Building an e-commerce business from scratch is no small undertaking. But experience has shown that a dedicated program properly structured can not only pull it off but do it far faster than many might think is possible.



Supporting your remote teams

As the COVID-19 pandemic spread, it overtook the effort of a retail chain to refine its new e-commerce site. The team, like workers around the world, had to learn quickly how to collaborate with one another while working remotely. The lessons from that experience and others include the following:

— Invest in adoption and standardization.

It’s not enough to have the best collaboration tools available. The level of people’s familiarity with collaboration tools varies, so spend time up front in training people how to use them well. Also standardize which tools to use. In an effort to move quickly, people may just gravitate to what they know best, resulting in the use of incompatible tools.

— Invest in adequate tools and support—for example, pay for upgraded Wi-Fi and distribute 4G or 5G modems.

— Stay committed to what works. Problemsolve the new issues—foreseen and unforeseen—that arise the way you do any other challenge: uncover the problems and innovate solutions by agile trial and error. Maintain frequent check-ins and track progress.

— Understand that in a time of crisis, team members have personal and family responsibilities that they must also handle. Have the empathy and flexibility to enable that.


Arun Arora is a partner in McKinsey’s Paris office, Philip Christiani is a senior partner in the Copenhagen office, Ralf Dreischmeier is a senior partner in the London office, Ari Libarikian is a senior partner in the New York office, and Hayk Yegoryan is a consultant in the Amsterdam office.

The authors wish to thank Karel Dörner for his contributions to this article.

Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.