8 Best Customer Acquisition Channels for DTC

The best acquisition channel for your store depends on your margin and repeat rate, not on anyone's ranking. A channel that pays for itself at a $200 average order loses money at $45, even when each customer costs exactly the same.

Below are the eight channels DTC brands put budget into: what a customer costs in each, how to set it up, and how to judge it against your own numbers.

Top 8 DTC acquisition channels at a glance

Email and SMS mostly sell to people another channel already paid to bring you. Marketplaces sell to the marketplace’s customer, and you rarely get the email address. Referral wins you real customers, but you set the price yourself, so there is nothing to benchmark it against.

All three take real budget and most brands count them as acquisition, so they stay in the list.

The example we use to compare channels

We use one made-up model store throughout: a Shopify brand with an $89 average order and a 38% contribution margin that won 1,000 new customers in one month. We follow those customers for a year, and the full breakdown is further down.

Margin per customer below is contribution over 365 days minus fully loaded cost. Positive means the channel covers what it cost to win the customer, before any share of fixed overhead.

Channels ranked by margin

The model store's five paid and organic channels, ranked by fully loaded cost and by margin per customer.

Affiliate buys the cheapest customer in the store and comes third on what that customer is worth, behind two channels that cost more to use. The two expensive channels stay at the bottom on both rankings. In this store, cheap and profitable mostly line up.

They don't always. The $40 example further down shows two channels at identical cost, one worth more than nine times as much as the other, and that gap is why the margin column decides and the cost column doesn't.

1. Paid social

Paid social is the main way to reach people who weren't looking for you at all. That's why it carries the largest share of the budget in most DTC brands, and the largest share of the losses.

The auction got more expensive this year. Meta reported average price per ad up 12% year over year in Q2 2026, with impressions up 14%.

Ad cost movement on both platforms in 2026.

Four habits decide whether the numbers coming out of the account mean anything.

• Split prospecting and retargeting into separate campaigns and report them apart. Retargeting your own buyers is retention spend wherever it sits.

• Exclude your customer list from prospecting audiences. It costs nothing and often turns up a third of spend going to people who'd already bought.

• Add new creative on a schedule, not after a bad week. Fatigue hits click-through rate first and cost second.

• Judge the account on new customers in Shopify. Ads Manager counts repeat buyers as purchases.

Your average cost hides what the next customer costs

Cost per customer is an average across everyone the channel delivered. A budget increase buys the next ones. They cost more, because the cheap audience went first.

Split the model store's $26,000 of paid social into three slices and the average stops being useful:

Media spend only, which is why these match the $70.27 and not the fully loaded $83.66.

Two things stand out. Every slice loses money, including the first $10,000, so on these numbers the channel is underwater on media alone before any retainer is added. And the losses get worse the further down you go.

Cut the last $6,000 and you save $6,000. You lose 60 customers worth $3,247 over a year, so the store ends up about $2,750 better off on cash for spending less.

The slices also assume the marginal customers repeat at the same 1.60 as the early ones, which is generous. Audiences you reach last usually repeat worse, so the real curve is steeper than this one.

All three slices of paid social cost more than a customer is worth, and each costs more than the last.

Your own curve comes from your last real budget change. Go and look at what happened to cost per new customer the last time you moved spend up or down. One warning: comparing two months also compares two promo calendars and two creative sets, so hold the offer steady or run it as a regional holdout.

2. Paid search and Shopping, non-brand

Non-brand search sells to people who already decided to buy the category. They are picking between sellers, so your price, reviews and delivery promise do most of the work before anyone reads your copy.

Alphabet's Q2 2026 filing shows paid clicks up 13% and cost per click up 3% across Google Search and its other properties.

Start by connecting the Google and YouTube sales channel in Shopify
This way stock and price sync into the feed on their own, because a feed carrying wrong availability spends real money on clicks to sold-out products. Keep brand and non-brand in separate campaigns while you're in there. Blended together, the cheap brand clicks flatter the account and hide what category terms really cost.

After that it's mostly hygiene.
Add negatives for the searches that describe a different product, since category terms pull in a lot of near misses, and send each click to the collection or product page that matches what was searched instead of the homepage.

Conversion rate is the strongest lever here. Cost per customer moves down as conversion rate moves up, and nothing about your offer or margin has to change.

Our benchmark of 1,000 Shopify stores found only 48% pass Core Web Vitals on mobile.

On Swag Golf, pages load three times faster. and mobile conversion rose 28%, which takes about 22% off cost per order on mobile traffic at identical spend.

3. Branded search

Branded search looks cheap in almost every account. In the model store it costs $39.68 fully loaded, second cheapest after affiliate.

It's also the channel most likely to be charging you for orders that were already on their way. Three studies, taken together, tell you how much of that risk applies to your store.

It comes down to who else is bidding on your name.

A Bing study of thousands of brands found brand ads lifted clicks by only 1% to 4% when nobody else was bidding. When competitors bid and the brand didn’t, they took 18% to 42% of the clicks.

Newer tests agree. Across Google branded search experiments in 2025, 82% of brands facing high competition measured significant lift, compared to 35% of those facing low pressure.

So the question isn't whether brand ads work in general. It's whether anyone is bidding on your name:

Search your own brand on a clean browser and note who sits above you. Then run the test: pause branded search across a set of regions for two weeks and watch total new customers in Shopify, not platform-reported conversions.

4. Organic search and content

In the model store, $8,000 a month of content and SEO wins 200 new customers, a $40 cost per customer. Most brands record it as free because no invoice arrives per click. It's also the only channel where the next customer gets cheaper as volume grows.

Counting it properly moves two numbers:

• Blended cost per customer goes from $52 to $60

• Blended LTV to CAC ratio goes from 1.2:1 down to 1.07:1

Work in this order:

• Collection and product pages for the terms people search once they've decided to buy. These are the ones that convert.

• The technical foundation underneath them, so the pages can be crawled and indexed at all.

• Comparison and buying guides for the research stage, with every piece linking to the products it recommends.

Organic and AI visibility are now the same job. Assistants read the same product pages Google does, covered further down.

5. Email and SMS

A welcome flow converts someone a paid ad or a search already brought to your site. The cost of winning that person belongs to whichever channel brought them.

The acquisition part of email is list growth, and it has a real price. A 15% first-order code at an $89 average order hands $13.35 to every new customer. That is money spent to acquire them whether or not it shows up in a media line.

Where email pays for itself is the second order.
That is what makes every other channel on this list affordable. Move the repeat purchase from month seven to month two and your 365-day LTV to CAC ratio doesn't change at all, but your cash position does, because the first cohort starts funding the next one.

Build the signup form first, with an offer worth an email address, and count what that offer costs. Then the welcome, browse abandonment and cart abandonment flows, all before any campaign goes out. The post-purchase flow comes next, timed to how often your product runs out, not to whatever suited the email calendar.

Keep the platform cost out of your paid CAC and report it against repeat revenue, or you'll make a retention channel look like an expensive acquisition one.

6. Affiliate and creator partnerships

Affiliate wins the cheapest customer in the model store, at $33.33, and the least loyal one. Those customers average 1.50 orders a year, so margin lands at $17.40, against $39.79 for branded search at a similar cost.

The commission explains it. $33.33 is about 37% of an $89 first order, and the store keeps only 38% as contribution, so the first order just breaks even and all the profit comes from the half order after it. Set commission against your own contribution margin, not competitors' rates. A rate that works at 55% margin loses money at 30%.

Shopify Collabs runs the program from your admin. It's free to install, with a 2.9% fee on commission payouts (about 97 cents on a $33.33 customer), on every plan except Starter and Retail. Shopify's Help Center says Collabs isn't accepting new creator signups, so plan to invite creators yourself rather than wait for new ones to find you.

Give every creator their own code or link, and watch for creator codes stacking on top of sitewide sales, which can push an order below cost.

7. Referral programs

A referral program lets you fix the cost per customer in dollars before a single sale happens. Affiliate rates are set in advance too, but they move with order value. A referral reward doesn't move at all.

It also has a ceiling nothing can lift, because a referral program can't produce more advocates than you have satisfied customers. Repeat rate is the usual proxy for that, though a one-off category with a genuinely loved product can run one well. What kills it is a product problem. Launch before that is fixed and you get a quiet dashboard and a wrong conclusion about referral.

Size the reward against contribution per order and count both sides of it, because two $15 rewards leave $3.82 of a $33.82 first order. Trigger the invite after delivery rather than at checkout, since nobody recommends a product they haven't received yet.

Then track referred customers as their own cohort. They usually repeat better than average, and that's the argument for raising the reward instead of trimming it.

8. Marketplaces

Amazon and TikTok Shop reach real demand you can't get to any other way, but they don't hand you a customer.

You rarely get the email address, so you can't bring that person back yourself. They may well buy again on Amazon, and none of it is a second order you can count on or influence. That second order is the entire argument for ever paying more than first-order contribution to win someone. Take it away and your ceiling drops to what a single order contributes.

How to treat it:

• Run it as its own P&L, judged on contribution per order after referral and fulfillment fees

• Keep marketplace orders out of your store's cost per customer and out of your store's repeat rate

• Compare the two on contribution, since one has a repeat curve and the other does not

Our guide to selling on Amazon and Shopify together covers how the fee stack works.

AI assistants, the channel everyone asks about

AI search gets written about more than any channel on this list and sends a fraction of the orders. We tracked AI referrals across 73 stores from September 2025 to September 2026, covering 96 million sessions and 977,000 orders in 2026 alone, to find out how many fewer.

• Visits ended the year where they started. Orders almost tripled. 896 visits a day in September 2025, 917 a day in September 2026, with a dip to between 640 and 680 through the winter and spring in between. Orders from those visits went from 5.9 a day to 16.1.

Orders from AI assistants reached 2.74 times their September 2025 level. Visits ended where they started.

• AI beat organic on conversion. On 41 of the 61 stores with enough volume, AI traffic converted better than Google organic. The median store saw a 25% higher conversion rate and 23% more revenue per visit, on orders 15% smaller.

• The share is still tiny. Across all 73 stores, AI assistants brought 0.21% of sessions and 0.26% of orders in 2026.

The work that captures it is product page work you should be doing anyway. Half of AI visits landed on a product page and converted at 1.77%. Homepage landings converted at 0.57%. So sizing, compatibility, price, stock and the returns policy belong in plain text on the page, where the assistant and the shopper can both read them.

Set up a custom channel group in GA4 so AI orders sit next to organic search in the monthly report. Traffic share on its own would have told those 73 stores that nothing changed in twelve months.

The AI channel has to sit above Referral, or GA4 files every AI visit as a referral and the channel stays empty.

Work out what each channel costs you

Producing your own numbers takes new customers by channel, not orders by channel. The two differ by every repeat buyer who clicked an ad.

In Shopify:

1. Open Analytics, then Reports, then Customer cohort analysis

2. Set Cohort definition to First order referring channel

3. Read the customer count per channel

4. Switch the metric to cumulative net sales and read down to month 11 for the 365-day figure, since month 0 is the acquisition month

5. Read orders per customer separately from the same cohort, since you need both

Cohort definition is the setting that gives you new customers per channel, which is the denominator cost per customer needs.

Spend goes in by hand. Meta campaigns never report cost, ROAS, CPA or CTR into Shopify. Google can, through the Google and YouTube channel app, but the columns often stay empty, so plan to pull spend from Ads Manager and Google Ads either way.

The Cohort definition filter needs Advanced or Plus. On Basic or Grow, export customer data and match first-order dates against UTM data in a sheet.

Your UTM hygiene decides whether any of this is true. Untagged links land in Direct. Mistagged ones land in the wrong channel. The channel you’re about to defund may be the one filing its customers correctly. Fix UTM standards before you read the table, not after.

The model store, split by channel

The store runs an $89 average order at 38% contribution margin, so every order contributes $33.82. In one month it won 1,000 new customers, 640 of them from a paid click, on $42,000 of Meta and Google spend plus a $5,000 agency retainer and $3,000 of creative production. Every spend figure here is that same month. The orders-per-customer column then follows that month's cohort for a year.

It is an illustration sized to a mid-market Shopify brand. No client data sits inside it.

Fully loaded cost spreads the retainer and creative across the three paid rows by share of media spend. Both scale roughly with spend, so leaving them out overstates your headroom.

The two channels carrying 90% of the media budget both lose money on every customer they bring. Paid social returns about 65 cents per dollar once fully loaded.

Covering acquisition is a lower bar than covering the business.
Carrying fixed overhead at 20% of revenue, this store needs a 2.1:1 ratio of contribution to cost before it makes money, which comes from contribution margin divided by margin minus overhead, or 0.38 / (0.38 - 0.20). We derive that in the LTV to CAC guide.

No channel you pay for clears it. Branded search comes closest at 2.0:1, organic reaches 1.8:1 and affiliate 1.5:1. Follow the cohort out a year and its $64,224 of contribution covers the $60,000 of acquisition with $4,224 left, against roughly $33,800 of overhead on the revenue it produced.

The model store is losing money, which is the point of using it. Your own channels can clear 1:1 and still leave the company short.

Direct is not a channel

Direct shows a $0 cost and the best margin in that table. It’s the line to distrust most.

Direct is the label analytics tools fall back on when they can't tell where someone came from. Some of it is word of mouth and typed URLs. The rest is:

• Untagged email links

• App clicks that drop the referrer

• A share of AI assistant traffic arriving without one

None of it grows because you decided to fund it. When Direct climbs, the first question is whether a campaign stopped tagging.

Rank by margin per customer, not by cost

Cost per customer only tells you what you paid, and nothing about who you got.

Take two channels both delivering customers at $40, on the same $33.82 of contribution per order:

Same cost, more than nine times the margin.

The gap is wider than most brands expect, because repeat buying is rarer than the industry admits. Bluecore's 2025 Customer Growth Benchmarks Report covers more than 100 retailers across seven verticals. Around three-quarters of retail customers buy once and never come back. Health and beauty retains new buyers better than any other vertical, and still loses most of them.

What happens to 100 new buyers over three years, from Bluecore's benchmark data.

Then read payback, channel by channel

Margin tells you whether a channel is profitable. Payback tells you how long your cash sits inside it. A brand can be profitable on paper in every channel and still run out of money.

Take cumulative contribution by month from the cohort report, filter to one channel, and find the month it crosses that channel's cost.

Two channels with the same 12-month contribution. The slow one ties up cash for six extra months.

Channels with a discounted first order sit at the slow end by design. Subscription and first-box offers price the first order at or below cost on purpose, so judge them at cycle three and cycle six, where subscription churn decides the outcome.

Which channels to add, and in what order

Channel choice is mostly a function of how much you already know about your own economics.

The last row is the most common one we walk into. Brands add channels to escape a cost problem when the real problem is that no channel has a number of its own.

Two rules we hold to on client accounts:

• Keep contribution margin above 30% before scaling paid. Below that, the first order rarely covers the customer.

• Build your contribution margin with returns taken out before you set any channel's ceiling. A refunded order never pays for the customer who placed it.

Your quarterly channel review

• Pull new customers by first-order referring channel for the last four mature cohorts

• Add fully loaded spend per channel, including retainers, creative and commissions

• Calculate contribution per customer minus fully loaded cost for each channel

• Plot cumulative contribution by month per channel and mark where it crosses that cost

• Check what your last real budget change did to marginal cost

• Confirm Direct did not grow because something stopped tagging

• Run one holdout test on the channel you trust most

Get a read on where your customers come from

Most channel problems we see aren't media problems. They trace back to a margin figure that was never right, a cost per customer built on orders instead of people, or a tracking setup that has been filing paid customers as organic for months.

Talk to one of our strategists if you want an outside read on what each of your channels is really worth.

FAQs about DTC customer acquisition channels