If you ask ten indie jewelry founders where their customers come from, eight will say "Instagram." If you then look at their actual revenue attribution — properly — you'll find that Instagram is rarely the channel doing the heavy lifting. The picture in 2026 is more interesting and considerably more useful.
The Channels, Honestly
Across the brands we've seen, customer acquisition cost (CAC) by channel typically falls in these ranges:
| Channel | Typical CAC | Best for |
|---|---|---|
| Email (existing list) | $15–$60 | Repeat purchase, new collection launches |
| Referral / word-of-mouth | $0–$80 | Bridal, fine jewelry, custom |
| Google Search (branded) | $20–$100 | Existing demand capture |
| Google Search (non-branded) | $200–$600 | Bridal, gemstone education |
| Instagram organic | Untracked, but real | Brand awareness, top-of-funnel |
| Instagram paid | $80–$300 | Fashion-jewelry, lower AOV |
| TikTok paid | $60–$250 | Fashion, gemstone-curious Gen Z |
| Pinterest paid | $50–$200 | Bridal inspiration |
The shape of these numbers tells a story. Email and referral dominate the cheap-CAC end. Search dominates the high-intent middle. Paid social is genuinely useful but rarely the lowest-cost channel for fine jewelry.
Why Email Keeps Winning
Email is the channel everyone forgets to credit, because it doesn't feel "active" — there's no daily algorithm to chase. But the math is unforgiving in email's favor:
- A list of 5,000 well-segmented buyers, sent a thoughtful monthly email with a 25% open rate and a 4% click rate, generates roughly 50 sessions per send.
- Those sessions convert at 4–8% (vs 1–2% for cold paid traffic), because they are pre-qualified buyers who know you.
- At a $400 AOV: 50 sessions × 5% conversion × $400 = $1,000 of revenue per send. Cost: ~$50 in tool fees and your time.
- Effective CAC: $1 to $5 per incremental sale.
Multiply across a year of monthly sends, plus automated welcome and abandoned-cart series, and a healthy email program is doing 25–40% of an indie brand's revenue with a tiny fraction of its marketing budget.
Why Instagram Is Misleading You
Instagram is genuinely valuable — but mostly for awareness, not direct attribution. The customer who eventually buys your $4,200 ring may have first seen you on Instagram nine months ago, then forgotten you, then googled you after a friend mentioned the brand at brunch. Instagram gets credit for nothing in your analytics. Google branded search gets all the credit. Both attributions are wrong; the truth is somewhere in the middle.
The trap: many founders see "Google branded search → $$$" in their reports and conclude they should spend more on Google. They should not. They should spend more on the channel creating the branded search demand, which is often Instagram, podcasts, press, or referral. This is the single most common attribution mistake we see.
What an Honest Channel Strategy Looks Like
For an indie brand under $3M, a healthy spend mix is roughly:
- 40–55% on email and CRM (tools, content, segmentation, automation).
- 15–25% on Instagram organic (content production, community, occasional paid amplification).
- 10–20% on Google Search (branded plus a small non-branded test budget).
- 5–15% on referral programs and PR (the channel everyone underinvests in).
- Rest on experiments (TikTok, Pinterest, podcast sponsorships).
Note what's missing: a giant Instagram-paid line item. For most fine jewelry under $3M revenue, paid Instagram doesn't pencil out — the AOV is high enough that the conversion math works on warm channels but rarely on cold ones.
What To Do This Week
Pull your last 90 days of orders. For each, assign the source as best you can — first-touch from your analytics, last-touch from UTMs, and self-reported survey data ("How did you hear about us?") if you collect it. Compare. The gaps tell you where your attribution is lying.
Then look at total spend by channel for the same 90 days. Calculate true CAC for each. The channel with the lowest CAC and the most room to grow is almost certainly email — and almost certainly underfunded. Fix that first.