Podcast: Download (Duration: 27:27 — 31.7MB)
The case for diversifying away from Amazon is that every channel you do not own can be taken from you without warning. During the pandemic Amazon refused shipments of non-essential goods, Facebook and Google both banned mask advertising, and Shopify shut down a seller for pricing hand sanitizer too high.
In this solo episode I recap what happened to both of my businesses in 2020 and what I am planning for the year ahead. My ecommerce revenue was cut by more than half in March, and the recovery came entirely through channels we control.
Below is all of it: what saved the business, why owned marketing is non-negotiable, how to increase customer lifetime value, and the iOS 14 change that is about to reshape online advertising.
Get My Free Mini Course On How To Start A Successful Ecommerce Store
If you are interested in starting an ecommerce business, I put together a comprehensive package of resources that will help you launch your own online store from complete scratch. Be sure to grab it before you leave!
Table of Contents
Key takeaways
- Revenue was cut by more than half in March 2020, and pivoting the product line recovered it.
- Owned channels saved the business. Facebook and Google both banned mask ads, so email and SMS carried the launch.
- Amazon refused non-essential shipments, which stranded anyone without their own fulfillment option.
- Even Shopify shuts sellers down. An acquaintance lost his store over hand sanitizer pricing.
- Ecommerce jumped from 11.8% to 14.3% of US retail sales in Q3 2020, per Statista.
- Repeat customers are over a third of our business, which is what makes annual growth stable.
- iOS 14 breaks cookie-based conversion tracking, so server-side APIs become necessary for advertising.
- Being early on Amazon best sellers invites attacks. Staying under the radar is now a deliberate strategy.
What happened to my ecommerce store in 2020
Revenue was sliced by more than half when the pandemic reached the United States in March.
The damage hit every product line simultaneously. Bumblebee Linens sells handkerchiefs for weddings and special occasions, plus personalized napkins for parties, hotels, and airlines.
Shelter in place removed all of that at once. Nobody was marrying, hosting parties, traveling, or staying in hotels.
A meaningful slice of the business serves real estate agents, and real estate sales stagnated through March and April too.
March 22 was the low point. My wife and I were seriously discussing moving to a cheaper location, since the warehouse runs close to $5,000 a month.
How we pivoted the product line during the pandemic
The recovery came from turning the same materials into products people actually wanted.
My wife’s idea was masks. We tested demand first by publishing a tutorial on converting one of our handkerchiefs into a face mask without sewing, and giving a free handkerchief with every purchase.
The promotion worked, so we imported masks and they sold quickly.
Quarantine gifts followed. The same monogrammed cocktail napkins, embroidered with pandemic jokes rather than wedding details, sold to people who could not see each other and wanted to send something.
Those two product lines carried us through April, which was the lowest point for morale the company has had.
Why owned marketing channels are not optional
Both Facebook and Google prohibited mask advertising during the pandemic, which meant paid channels were unavailable exactly when we needed them.
Email and SMS are what launched that product line. Without our own customer list, the recovery would not have happened.
Amazon compounded the problem by suspending non-essential shipments to its warehouses, which prevented restocking sold-out inventory.
The general principle is that anything you do not own can be withdrawn. Shopify shut down an acquaintance selling hand sanitizer because they judged his prices too high, when his own costs had risen and he was pricing to stay profitable.
If Shopify can decide what you are permitted to sell, no third party platform is a safe foundation.
Why Amazon sales are the most stressful revenue you have
Amazon sales grow every year for us and remain the least satisfying and most stressful part of the business.
The failure modes arrive without warning. You wake up to suspended listings or a changed listing, and support can take days to respond at all.
Malicious seller activity has not improved. In episode 332, Amanda Wittenborn described having her most profitable listings taken down for copyright infringement on products she designed and owns the copyright to.
Her current strategy is telling. Release many products and deliberately stay under the radar, because hitting the best sellers list is what makes a listing a target.
Having to avoid becoming a best seller is an unusual position for a platform to put its sellers in.
The inventory problem with relying on FBA
Amazon’s inventory limits do not account for the holiday surge, which strands sellers at the worst time.
Warehouse space was genuinely scarce over the 2020 holidays, on top of the earlier refusal to accept non-essential shipments.
We ran out of stock and were forced to go merchant fulfilled, which was only possible because we have our own fulfillment operation.
A seller without that capability and without their own sales channel had no options at all during that period.
Why 2021 is a good time to start an ecommerce business
Ecommerce as a share of US retail sales rose from 11.8% to 14.3% in Q3 2020, according to Statista.
That figure is more striking than it looks, because Q3 came after stores had reopened.
The behavior change appears permanent. My own mother resisted online shopping for years, started during the pandemic because she did not want to go to the grocery store, and has no intention of going back.
The important qualifier is which kind of ecommerce. Marketplaces will get more competitive as more sellers arrive, which strengthens rather than weakens the case for owning your own platform.
How to increase customer lifetime value
Segment your best customers and treat them differently from everyone else.
We review our customer list quarterly to identify people who buy frequently and in volume. Those customers get early product releases, special coupon codes, or a dedicated representative for their orders.
Email cadence matters more than most sellers think. Sending less than once a week means leaving sales on the table.
The core automated flows should all be running: pre-purchase, abandoned cart, post-purchase, and win back.
Average order value is the other lever. Bundling complementary items and raising the free shipping threshold both push orders upward.
How to cross-sell based on what someone bought
Route customers into different post-purchase sequences depending on their first purchase.
We rebuilt this in Klaviyo at the start of the year. Someone buying cocktail napkins enters a sequence that cross-sells dinner and lunch napkins specifically.
Facebook Dynamic Cross-Sells handles the paid side automatically. Facebook knows what products your customers buy together and generates ads for the likely next purchase.
A loyalty program adds another retention layer, and ours runs through ManyChat at no cost.
The result is that repeat customers are now over a third of our business, which gives every year a stable base to grow from.
What the iOS 14 tracking prompt does to advertising
Apple now prompts users for permission to track them across apps and websites owned by other companies.
The prompt is obstructive and the wording invites refusal, so most users are likely to opt out.
Opting out removes your visibility entirely for that user. You cannot tell whether they visited, viewed a product, added to cart, or purchased.
Apple users are a large share of devices, so losing their conversion data degrades ad effectiveness broadly and makes your audiences inaccurate.
The fix is server-side tracking. Rather than relying on browser cookies, your server transmits conversion data directly to Facebook or Google through their conversion API.
Shopping cart support for these APIs is arriving, and you need to implement it separately for each advertising platform you use.
What to do if you only sell on Amazon
Start building your own site now, accepting that initial sales will be lower than what Amazon produces.
The timeline is the reason to begin immediately. Building your own traffic takes considerably longer to ramp, so starting after something goes wrong is too late.
Something eventually will go wrong with your Amazon account, which is close to the only guarantee in this business.
Focus the effort on channels you own. Most sellers have email running and far fewer have SMS, which is currently performing well for our store.
Selling across many marketplaces is fine. Relying on any one of them for the bulk of your revenue is what creates the exposure.
How to start when you do not know what to sell
Publish content on something you are interested in and let the audience tell you what to build.
The startup cost is close to nothing. A YouTube channel is free, a podcast can start around five dollars, and a blog costs a few dollars a month.
Content is also the durable traffic strategy regardless of advertising conditions. It attracts people you can later convert into customers.
The skills that determine whether this works require no expertise: showing up on schedule, putting real effort in, staying genuinely interested, and maintaining a consistent work ethic.
Being coachable matters more than most people expect. When someone points out that you are doing something wrong, examining it beats explaining why it cannot be helped.
Why you have to give a business three to five years
Expecting success in year one is what causes people to quit before anything compounds.
My reference point is a college engineering course, the hardest I ever took, where I understood nothing for an entire semester and had it finally land in the last two weeks.
Business works the same way. Effort accumulates invisibly for a long stretch before results appear.
The attitude that stops progress is claiming to have tried everything. If you had genuinely tried everything, one of those things would have worked.
Frequently asked questions
Why should you not rely only on Amazon?
Because everything about the channel is outside your control. During the pandemic Amazon refused non-essential shipments, holiday inventory limits stranded sellers, and malicious competitors can get profitable listings suspended with little recourse.
What are owned marketing channels?
Email, SMS, and to a degree Facebook Messenger. These are lists you control rather than platforms that can ban your category. When Facebook and Google both banned mask ads, our email and SMS lists are what launched the product.
Can Shopify shut down your store?
Yes. An acquaintance selling hand sanitizer was shut down because Shopify judged his prices too high, when his own supplier costs had risen and he was pricing to remain profitable.
How do you increase customer lifetime value?
Segment your best customers and give them early releases, special codes, or dedicated service. Email at least weekly, run the standard automated flows, and cross-sell based on what each customer actually bought.
What does iOS 14 change for Facebook advertising?
Apple prompts users to allow tracking across apps and websites, and users who decline become invisible to your pixel. You lose visibility into their visits, product views, cart adds, and purchases, which degrades both reporting and audience quality.
How do you fix conversion tracking after iOS 14?
Move to server-side tracking through the platform’s conversion API, where your server sends conversion data directly rather than relying on browser cookies. You implement it separately for each advertising platform.
Is 2021 a good time to start an ecommerce business?
Ecommerce rose from 11.8% to 14.3% of US retail sales in Q3 2020 even after stores reopened, and the shopping behavior change looks permanent. The qualifier is building on a platform you own rather than a marketplace alone.
How long should you give a new business before expecting results?
Three to five years. Effort accumulates invisibly for a long stretch before results show up, which is why expecting first-year success is the most common reason people quit too early.
