TLDR: If you're attributing Shopify sales to a webinar (or any campaign), you MUST exclude orders tagged Subscription Recurring Order. They're Recharge auto-rebills. Your event didn't cause them. Keep Subscription First Order — that one's real.

The Metric I Was Building

For our supplement brand's webinar analytics dashboard, I track something I call the Halo Effect — registrants who purchased without an event discount code.

The logic: if someone attended the webinar and later bought without using a promo code, the event probably influenced them even if we can't directly attribute it.

It's a meaningful signal when the data is clean.

The First Fix That Wasn't

My first instinct when the numbers looked off was to tighten the time window.

Maybe I was pulling in sales from weeks after the event. So I scoped the Halo Effect down to the actual event window only.

That helped. But the totals still felt inflated.

What Was Actually In There

Recharge (Shopify's subscription billing integration) auto-rebills existing subscribers on a set schedule. When someone's renewal happens to land during the event window… it passes the date filter.

And there it is — a "sale" that had absolutely nothing to do with our webinar.

The Shopify orders export has a Tags column. Recharge stamps every order with one of two values:

  • Subscription Recurring Order — an automatic rebill of an existing subscription. Not event-driven. Exclude it.
  • Subscription First Order — the first order of a brand-new subscription. A genuine new purchase. Keep it.

The trap is assuming you can just drop all subscription orders. You can't.

A first-time subscriber who found us through the webinar and signed up is EXACTLY the win you're trying to measure. Nuance matters here.

The Fix

In shopify-all.ts, I parse the Tags column (lowercased). In attribution.ts, I skip any order carrying the subscription recurring order tag before it enters the halo loop.

One filter. Big difference.

What the Real Numbers Were

On the reference event: 5 of 34 halo orders were Recharge rebills.

After excluding them: 34 orders → 29.

That's not noise. And if I had been using the inflated numbers to decide which affiliates to re-invite or whether this webinar format was worth repeating — I'd have been optimizing against fiction.

Why This Matters to Me

Accurate metrics aren't just a nice-to-have.

They're how you make the call on which offer to run again, whether the Halo is growing, whether the webinar is actually moving buyers who wouldn't have bought anyway.

If your Halo Effect is padded with renewals that Recharge was going to process regardless of your event, you're measuring your subscription product's billing cycle — not your marketing.

Check your Tags column. The signal is right there.