195: The Software Tools I Use To Run My 7 Figure Online Store With Steve Chou

Podcast: Download (Duration: 34:39 — 47.9MB)

Email drove over 35% of Bumblebee Linens revenue this past holiday season, which is the clearest argument there is for selling on your own site rather than relying solely on a marketplace. That 35% came from repeat customers and people who already knew the brand, some of whom have bought from us for ten years.

I get between 100 and 400 emails a day, and a large share ask which tools I actually use to run the store. This is the answer, covering the shopping cart, email platform, popups, push notifications, shipping, and payment processing.

A note on timing: this stack is from 2018 and several products have changed or disappeared since. The selection criteria still hold even where a specific tool has moved on.

Key takeaways

• Email produced over 35% of holiday revenue, almost all from repeat customers and existing brand awareness.
• Avoid cheap site builders for ecommerce. Their limits force a painful platform migration later.
• Owning your source code lets you build any feature you need, which is worth real money on conversion fixes.
• A virtual private server running several sites costs around $100 a month versus $500 to $600 for equivalent hosted functionality.
• Segment promotions to exclude recent buyers of the discounted product, so nobody sees a sale on what they just paid full price for.
• Push notifications work like email for browse abandonment, reaching both phone and desktop.
• Run a backup payment processor, since primary processors do go down during heavy traffic.
• Calling customers who abandoned carts revealed a coupon-entry problem no analytics tool would have surfaced.

Which shopping cart should you use?

Choose based on how much you want to touch code. If you are tech averse, go with a fully hosted platform like Shopify or BigCommerce and stop thinking about it.

I run a heavily modified open source cart I have maintained since 2007, and I do not recommend anyone follow that path today. Better options exist now that did not exist then.

If you do want to own your source code, OpenCart is the one I would choose starting over. It is fast, the code is readable, it is feature rich, and it has a real plugin library behind it.

Why you should not use a cheap website builder for ecommerce

General-purpose site builders offer enough ecommerce functionality for a small shop and run out of room exactly when you start succeeding. That is the wrong time to discover a platform limit.

The cost is the migration. Switching carts once you have real traffic, real SKUs, and real order history is a significant undertaking, and it is one of the main reasons I have never moved off my own cart.

Judging the right platform before you have sold anything is genuinely hard. The practical rule is to avoid being cheap about the one piece of infrastructure everything else sits on.

Should you build an ecommerce store on WordPress?

WordPress is a content management system rather than an ecommerce platform, and running a store on it means carrying plugins that break on upgrade. Every upgrade with money on the line is a gamble.

Speed is the second problem. WordPress carries code an ecommerce site does not need, which is why so many WordPress stores are slow and need more powerful hosting to compensate.

I have watched students lose patience with things breaking every six months, and I have critiqued sites that were slow purely from plugin load and missing caching.

If you are committed to WordPress, Ecwid is the option I like. The store itself runs on their servers and embeds into your site, so the ecommerce load never touches your WordPress install.

What does hosting an ecommerce store actually cost?

A virtual private server running four or five sites costs about $100 a month and is nowhere near capacity. The equivalent functionality on a fully hosted platform would run $500 to $600 a month.

That gap is the tradeoff for managing it yourself. You get full control and you own the maintenance.

Why owning your source code matters

Owning the code means any feature you need is a coding problem rather than a feature request. I have used that repeatedly for things no platform would have offered.

One example is custom sales reports built to my wife’s exact specifications for accounting, written in an afternoon.

The more valuable one came out of a real conversion problem. During a holiday apron promotion we had unexplained cart abandonment, so I called the people who abandoned.

They had received a coupon code, reached checkout, forgotten the code, hit back, could not find it again, and left frustrated. No analytics tool would have told me that.

The fix was automatic coupon application through a cookie, a few lines of code. I extended it so that anyone who trades an email for a coupon has it applied automatically at checkout, valid 24 hours with a countdown timer for urgency.

Why use an ecommerce-specific email platform?

An ecommerce email platform tracks every customer and every sale, which lets you build segments on purchase behavior rather than on tags you set up in advance. That is what makes revenue-maximizing promotions possible.

The straightforward example is coupon targeting. Long-time newsletter subscribers who have never purchased get a coupon, and the same list members who bought recently do not, handled dynamically.

Our 12 Days of Christmas promotion is the sharper version. Each day carried a flash sale on a specific product, and every send excluded anyone who had bought that exact product in the prior two weeks.

That exclusion matters more than it sounds. Nothing sours a good customer faster than a 50% off email for something they paid full price for last week.

These platforms also push custom audiences to Facebook automatically, so the same segment can receive matching ads.

Cost is real and the return is clear. For a store built on repeat business rather than one-time transactions, it earns out.

How do you build better email capture forms?

A dedicated popup tool beats hand-coding once the designs get complex. I build most things myself and stopped doing it for popups.

The change came from wanting a spin-to-win wheel, where visitors trade an email for a chance at a prize. No out-of-the-box version existed for my cart at the time, and the tool I used was flexible enough to let me build one.

That single form more than doubled email signups.

The wheel is now a standard feature in most popup tools, alongside drag-and-drop builders. I now use one for everything simply because designing there is faster than writing the HTML.

How do push notifications work for ecommerce?

Push notifications reach both phone and desktop after a visitor opts in, and an ecommerce-specific tool can trigger them off your product catalog automatically. Think of it as email for people who never gave you an email address.

The behavior is the same as email automation. Someone browsing a specific product without buying gets a reminder four hours later and again two days later.

Post-purchase works the same way. A customer who has not bought again within a month receives a notification with a coupon.

Flash sales are where it performed best. A 50% off notification appears on the phone, one tap reaches the product, and checkout follows immediately, which suits impulse buying well.

What shipping software should you use?

Pick one with a free tier so you can adopt it before the volume justifies paying. That is the entire reason I use ShippingEasy over ShipStation or ShipWorks.

The free tier lasts until you exceed a volume threshold, which for a new store can be a long time. By the time we crossed it, the workflow was embedded.

Feature-wise the major options are comparable and all integrate with the popular carts. The free entry point is the differentiator when you are starting.

Why run three payment processors?

Because processors go down, and during heavy traffic that is expensive. Over an eighteen-month period our primary processor failed multiple times, including during peak periods.

PayPal is effectively required for mobile despite being among the more expensive options. It imports customer details so nobody types an address or card number on a phone, which materially improves mobile conversion.

A second processor sits behind it as a backup. When the primary fails, transactions continue rather than stopping.

The redundancy is paranoia and it has paid for itself. A processor outage on a high-traffic day costs more than running a spare account.

Why selling on your own site beats relying only on a marketplace

Your own store gives you repeat sales and an actual customer relationship, which a marketplace does not. The 35% of holiday revenue that came through email is the concrete form of that.

Those buyers include people who have purchased from us for ten years. No marketplace hands you that relationship or lets you email it.

Frequently asked questions

What shopping cart should a beginner use?

A fully hosted platform like Shopify or BigCommerce if you do not want to deal with code. If you want to own your source, OpenCart is fast, readable, and has a real plugin ecosystem.

Can you use Wix or Squarespace for an online store?

They work for a small shop and their ecommerce features are limited. You will likely outgrow them and face a difficult migration exactly when the business starts working.

Is WordPress good for ecommerce?

Not especially. It is a content management system at heart, plugin-heavy stores break on upgrades, and the extra code slows sites down. An embedded store widget avoids most of that if you must use WordPress.

How much does it cost to host an online store?

Around $100 a month for a virtual private server running several sites, versus $500 to $600 for comparable functionality on a fully hosted platform.

How much revenue can email drive for an ecommerce store?

Over 35% during our holiday season, almost entirely from repeat customers and people already familiar with the brand.

Should you exclude recent buyers from promotional emails?

Yes. Sending a 50% off email for a product someone bought at full price two weeks ago damages the relationship, and purchase-aware platforms exclude them automatically.

Do push notifications work for ecommerce?

Yes, for browse abandonment, repeat purchase reminders, and flash sales. They reach both phone and desktop and suit impulse purchases because checkout is one tap away.

Do you need more than one payment processor?

A backup is worth having. Primary processors go down during heavy traffic, and an outage on a peak day costs more than maintaining a second account.