I’ve done it thrice, first with Flipkart, then with a company that I founded myself, then at Amazon. I still have PTSD.
But e-commerce is big and is growing like gangbusters. U.S. consumers are estimated to have spent $933.30 billion on e-commerce in 2021, up 17.9% YoY. This was after a bumper 2020 when the industry grew 32.4% YoY, pulling forward five years of projected growth into one year.
There’s no turning back now. In 2022, one out of every six dollars will be spent in the online, up from about one out of ten in 2019. Businesses that don’t have an online store are finally getting in the game.
Several e-commerce platforms such as Shopify and WooCommerce have made building an online store easier than ever before¹. But even as demand is moving online, sustaining this demand continues to become more challenging. Buyers now have hundreds of purchase options just a click away and their attention is more constrained by mounting distractions, stress, and social needs.
This guide breaks down the four key business levers that are in your control to grow your e-commerce business: visitors, conversion rate, order value, and number of orders (shown as the four major quadrants in the revenue map below). In the next few sections, we further break down each of these four strategic levers into key metrics that you can use to develop experimentation themes for your teams and organization.
Converting visitors who’re already at your store is one of the highest impact levers in e-commerce. Entire companies and agencies are sometimes devoted only to conversion rate optimization (CRO). The problem with applying such tactical CRO solutions is that most optimizations in one part of user experience often come at the cost of more important downstream impact on user experience. Without an understanding of such downstream impact, most conversion optimizations turn out to be shots in the dark.
“At Stitch Fix, we actually had a team of data scientists that were just focused on performance marketing. Around them was the larger team of marketers, product people, and finance. The aim was to align cross-functional groups around business levers. The broader trend of optimizing for downstream client value versus upfront acquisition cost was one of the biggest shifts during my time there.” Mike Duboe, former Head of Growth, Stitch Fix
Statsig’s unique Pulse metrics bring to you the whole business picture so that every team member can now read the full impact of an experiment, from landing page views to product detail views, to cart conversion and purchase events in one place (see below).
Tip: If you must choose one step to optimize, pick conversion steps deeper in the funnel before moving to upstream steps. Large improvements deeper in the funnel require a smaller sample size to test and make every upstream step more effective.
Nearly 7 out of 10 buyers are known to abandon their carts². If the top reason for cart abandonment is shipping costs, offering upfront clarity may be a major low hanging and a high impact area for experimentation.
Here are a few more tactical options to reduce cart abandonment that I’ve tried:
Streamline checkout into a single page: Single page checkouts enable buyers to see and complete sections faster and with more confidence. For example, an A/B test for checkout on the Vancouver Olympic Store showed that a single page checkout performed 21.8% better than the multi-step checkout.
Add guest checkout: Making checkout as frictionless as possible with guest check out or popular sign-in options from Google/Facebook/Apple could be single biggest UX improvement you can make to improve purchase conversions.
Earn trust with upfront clarity: Here’s Wayfair with clear information on the free shipping threshold and estimated delivery dates presumably based on the zipcode information from my account.
Provide flexibility with more payment options: According to a Dec 2021 study, over half of buyers want more payment options at checkout. Again, here’s Wayfair again with no less than six payment options in addition to the default credit card option! While buyers want more options, multiple tests show that payment selection can introduce complexity and choice paralysis. Optimizing payment options has been been one of the most fertile grounds for experimentation at Amazon (examples include options from Gift Cards to Debit Cards to an in-store balance).
If you want to dive into A/B testing your checkout flow, here are more ideas to grow your purchase conversion rate.
Today, the buyer’s path to purchase is a winding journey that may begin on Google, followed by noticing the product on the shelf of a physical store, researching reviews and pricing on Amazon, and hitting the Buy button on an Instagram feed. Meeting buyers where they are means that e-commerce businesses now increasingly employ a multi-channel strategy. This makes it critical for you to separate the forest from the trees.
Site-wide add-to-cart conversion rates may offer little-to-no signal for you to optimize for buyer journeys spanning different channels. With Statsig, you can add dimensions to capture a buyer’s past and current actions across different channels. These dimensions allow you to breakdown your key metrics such as total product_viewsand purchases by the channel dimension, and separate which channels are more effective with your buyers. Specifically, when you run experiments to improve conversion rates, you can breakdown the conversion rate results by channel to find where your experiment had the most impact.
Statsig’s Autotune is another powerful way to optimize add-to-cart conversion rates. Autotune automatically picks the optimal product hero image from a set of variants to that leads to the most downstream purchases.
Driving traffic to your store is a constant marketing challenge. What worked last month may not work next month.
In terms of marketing and where you show up, I think one of the biggest things that we’ve learned is that the world changes so quickly. When we launched three and a half years ago, the things that were working for us are now things that everyone’s doing, so you really have to stay on top of it. We’re constantly looking for new ways to pop up wherever our clients are to make sure that it’s relevant to them and it doesn’t just get lost in the shuffle. Changing up creative constantly and making it new and different. Cofounder and President of M. Gemi, Cheryl Kaplan
While the flavor of the month may constantly change, few things trump creative content in terms of return on investment. Creative content remains ahead of reach, brand, targeting, and context in terms of marketing effectiveness³. But how to measure content effectiveness?
Effective content grabs and holds buyer attention, offers a path to explore the product further, and clearly sets up the final desired action for the user.
More than 1 out of 3 users who search for a product on Google tend to make a purchase within five days⁴. Optimizing your site’s content for relevant keywords for SEO (Search Engine Optimization) pays dual dividends: increasing your store’s relevance to visitors who land on your site and increasing your store’s reach among prospective buyers. Effective content can also result in better site performance, which results in higher search ranking. For example, in 2017 Airbnb ran a search result landing page experiment, internally known as the “Magic Carpet”, that featured a large header with an image and search box, along with extra content such as reviews & listings beneath (see Test Treatment below).
We hypothesized that this landing page would increase relevance with its clearer content, and decrease page load time with its lighter code structure, among many other improvements. This would consequently lead to a higher ranking for our page on search engine results.
While content remains central to extending reach, Nielsen suggests that modern data analytics have improved the effectiveness of targeting, reach, and recency that now contribute 36% to advertising effectiveness (up from 15% over the past 11 years). My takeaway from this is that when a feature or campaign isn’t well targeted to buyers, content must work even harder to drive sales.
As we mentioned in Essential Growth Frameworks, if you’re trying to expand reach with paid search or paid social channels, it should be purely incremental to your business:
Any new paid acquisition channel should be incremental to your current acquisition channels. It should help acquire customers that you would not have gotten anyway. Measure the marginal RoI, not the absolute RoI.
Statsig enables you to build sophisticated targeting mechanisms with Segments. You can create segments using ID lists as well as conditional rules, and reuse these segments for progressive feature development and experimentation.
Referral programs are a popular way to target buyers through their trusted personal network. Tesla has experimented with long list of referral programs. Here’s Casper offering buyers a $75 Amazon Gift Card for referring a friend who gets 20% off on their order. They even have what I’d call an ‘experimental’ referral program… get $10 off for referring a friend to a pillow!
To hit the punchline of this section: you may not find a single customer that places orders equal to the AOV of your business. To grow your AOV, it’s more valuable to breakdown your order values into buckets and grow the most common order value, aka the modal (not avg) order value.
Here are a few ideas to grow the modal order value:
Product recommendations are one of the most straightforward ways to grow order value. Enabling wish lists and promotional offers based on browsing history are simple ways to ‘recommend’ products to your buyers based on their past intentions. Recommendations based on buyer behavior can also take the form of people who bought this item also bought this or appear as store-recommended styles with complete the outfit and pair with options as shown below.
Discounts for product bundles incentivize buyers to buy more of what they like or to buy an all-in-one solution for their needs. Who doesn’t need more socks, especially when you keeping losing one of each?
Want to save your relationship? Better get the Relationship Saver!
No one should be stuck with the same coffee every month, no matter how good it is. Trade Coffee combines personalized coffee matching and discovery with subscriptions. All coffee growers should make a decent livelihood.
Setting thresholds for free shipping and promotional offers for “all orders above” set buyers on the path to making the most out of their purchase. Glossier blows it out of the park on both, setting up two thresholds of order values.
Deal countdowns and limited period/edition offers create a sense of scarcity, converting mild desire into immediate action. Only Amazon can get away with a ‘countdown deal’ of $1.99 instead of $2.99. But if you’re looking for a bed, Ikea is offering the Malm bed for $249 until Jan 31 so you can sleep easier. No rush.
Loyalty programs such as Costco Memberships, Amazon Prime, and airline loyalty programs are household names and subjects of case studies for growing the frequency of purchases and customer lifetime value (LTV). For example, Prime members are known to spend more than twice than non-prime members on Amazon.
‘It might make sense for stores with a large number of SKUs to offer loyalty programs. But how do smaller companies employ loyalty to increase number of orders and total buyer spend?’, you ask? The short answer is building a relationship with your buyers regardless of the size of your product portfolio pays handsome dividends… For example, Prose more than tripled their sales in 2020 driven by subscriptions that launched in Mar 2020. Together with an exclusive loyalty program that launched in July, members enjoy Prose as the one-stop-shop for all their hair care needs with exclusive consultations and perks including 15% off orders.
“2020 was a foundational year in establishing our membership program and we’ve been analyzing our data to ensure any new incentives are truly valuable and helpful for our customers.” Prose co-founder and VP of Product, Paul Michaux
Mizzen+Main offers a tiered reward program with “best damn rewards possible” including gifts, invites to exclusive events, and a personal concierge.
Personalization and exclusive events don’t cut it for your buyers? Try other incentives… Diamond Candles’ RingReveal has got buyers hooked and coming back for more with surprise fashion rings included in their candles. The chance to win a real diamond ring can’t hurt either!
While conversion rates, visitors, AOV, and number of orders may each serve as the primary metric for different teams with different goals, serving the broader business objective requires each team to track business-wide metrics to avoid the unintended effects of tactical or isolated efforts. Together, these metrics form a shared business vocabulary for data-driven decision making and provide the unifying context for each team to make the better, faster decisions everyday.
If you’re looking to run your first experiment, want advice on an experiment, or simply want to join the conversion, hop over to Statsig’s Slack channel to meet hundreds of others on a mission to power growth for their businesses.
Thanks to our support team, our customers can feel like Statsig is a part of their org and not just a software vendor. We want our customers to know that we're here for them.
Migrating experimentation platforms is a chance to cleanse tech debt, streamline workflows, define ownership, promote democratization of testing, educate teams, and more.
Calculating the right sample size means balancing the level of precision desired, the anticipated effect size, the statistical power of the experiment, and more.
The term 'recency bias' has been all over the statistics and data analysis world, stealthily skewing our interpretation of patterns and trends.
A lot has changed in the past year. New hires, new products, and a new office (or two!) GB Lee tells the tale alongside pictures and illustrations:
A deep dive into CUPED: Why it was invented, how it works, and how to use CUPED to run experiments faster and with less bias.
Explore Statsig’s smart feature gates with built-in A/B tests, or create an account instantly and start optimizing your web and mobile applications. You can also schedule a live demo or chat with us to design a custom package for your business.