Ecommerce CRM in 2026: Pick by the Problem That Costs You Most, Not by the Feature List

An online store rarely loses money because the product is bad. It loses money
because the same customer is a stranger in three different systems: a buyer in
the order table, a contact in the mailing list, and a ticket in the support
inbox. Every campaign, flow and reply built on top of those fragments misses
context it should have had. An ecommerce CRM exists to merge the three into one
profile, so that what the customer actually bought drives what happens next.

This piece is a companion to a longer comparison on our blog that walks through
six tools with their September 2026 prices. Here I take a different axis: start
from the failure mode that hurts your store today, and let that decide the
category first. The tool shortlist comes after.

Failure mode one: buyers who never come back

The symptom is a store with decent traffic and a dead mailing list. First-time
buyers check out once and vanish. One-category buyers never discover the second
category. High-value customers get the same generic newsletter as everyone else.

What fixes this is marketing automation built around store events: abandoned
cart, browse abandonment, post-purchase follow-up, review requests, and win-back
after a couple of months of silence. The trigger is an event from the shop, not
a manual campaign you remember to send. Segments based on purchase behaviour
should update themselves as orders arrive, without anyone rebuilding a list
every week.

In this corner sit the marketing CRMs built around email, SMS and behavioural
automation — Klaviyo, Omnisend, Mailchimp and MailerLite in the comparison. The
default for Shopify and direct-to-consumer brands is Klaviyo, because its flows
are built around store events and its segmentation runs deep. The entry paid
Email plan is twenty dollars a month for a few hundred active profiles, and the
Email plus SMS plan is thirty-five a month for the same band. The free plan
covers up to two hundred and fifty active profiles with a few hundred sends,
which is enough to validate flows before paying anything.

The value pick for a small shop is Omnisend: same ground — email, SMS, web push,
automation — with a simpler interface and a free tier of two hundred and fifty
contacts. Paid plans start at sixteen a month for Standard and fifty-nine for
Pro, with an introductory discount for the first three months that you should
not budget around, because the full price returns after.

MailerLite is the budget option that still supports real automation, with a
free plan and paid plans starting in the low teens per month. Its ecommerce
features are lighter — no deep product-recommendation engine — so it fits when
email automation is all you need and the difference goes to ads instead.

Mailchimp remains reasonable when you want email, landing pages, forms and a
light CRM in one account, with one change that matters for 2026: the permanent
free plan is gone. What remains is a trial of about two weeks, after which the
paid tiers start in the low teens per month, and the cart recovery and product
recommendation features sit in the higher tiers.

The trap shared by this whole category is the billing unit. Klaviyo bills on
active profiles, not on emails sent — a list full of people who never open
anything still costs money every month, so list hygiene is part of the budget.
Omnisend and MailerLite price on contacts and monthly sends. Cleaning the list
is the cheapest optimisation available in this category.

Failure mode two: support tickets eating the day

The symptom is different: sales are fine, but "where is my order", returns and
refunds eat the working day, or pre-purchase questions pile up on items that
need explanation. A generic helpdesk without an order sync turns every ticket
into a research task, because the agent cannot see the order, the tracking
number and the previous tickets on the same screen.

What fixes this is a helpdesk built for ecommerce, where every ticket carries
the order and its history. Gorgias is the representative in the comparison: it
is priced by ticket volume, never per agent, with an entry plan around forty a
month plus a helpdesk add-on, scaling through ninety, five hundred and beyond
as ticket volume grows, with a per-ticket fee past each limit. The AI agent is
a separate line on the invoice, from thirty a month on the entry plan upward,
plus a fee for every automated interaction past the allowance — useful, but
budget it as its own line.

The rule of thumb: below roughly fifty tickets a month, email or a shared inbox
is enough. Above three hundred a month, a helpdesk that shows the order inside
the ticket usually pays for itself, because it removes the manual lookup step
from every conversation.

Failure mode three: quotes that go cold

Wholesale, quote-based and business-to-business sellers have a third problem:
their "orders" start as quotes and follow-ups, and a marketing automation tool
will not send abandoned-cart flows for a pipeline that lives in spreadsheets.
What fixes this is a cheap pipeline CRM run alongside a marketing tool. Zoho
Bigin is the example: free for a single user with a few hundred records, then
paid tiers in the single digits and teens per user per month. It will not send
cart flows — that is not its job — but it keeps wholesale buyers visible where
a marketing platform cannot.

The invoice is always higher than the headline

Four items explain why the final invoice usually lands well above the advertised
entry price. Contacts versus sends, as above. SMS and push as separate credits:
an email-plus-SMS plan includes a small monthly allowance and every extra
message is billed, with SMS costing far more per send than email. Per-ticket
overage in helpdesks: a single viral product launch can push a store into
overage for a month. And AI add-ons billed apart from the base plan.

The marketplace wrinkle in Southeast Asia

Stores in Vietnam and the region rarely live on a single storefront. Orders
arrive from Shopee, Lazada, TikTok Shop, Zalo OA and the store's own website.
Global CRMs connect natively to Shopify and WooCommerce, not to the local
marketplaces — so in practice a shop either consolidates orders into a local
platform first, or exports orders into the CRM on a schedule and accepts that
segmentation runs a day behind. A workable stack is a local sales and inventory
platform for orders and stock, Zalo OA for conversations, and one marketing CRM
for retention flows.

Two compliance points belong in the same decision. Vietnam's Personal Data
Protection Law has been in force since the start of 2026, with its decree
setting out consent records, impact assessments and breach notification within
three days. And since the middle of 2026, electronic invoices are required for
household businesses as well as companies, so an invoice number should be part
of the order record the CRM keeps. Consent logs and invoice numbers tied to
order records are evidence a shop may have to produce on demand — keep them
exportable, not locked inside a single rented account.

Decide in three questions

First, where do you lose money today: customers who never return, unanswered
tickets and returns, or quotes that go cold. That picks the category. Second,
how many contacts do you carry and how many are buyers: under a few thousand
engaged contacts the entry plans are enough, and above that the pricing
calculator decides the tool. Third, do you want support and marketing in one
login: one tool is cheaper and simpler, while two specialised tools usually
work better once support volume passes a few hundred tickets a month.

The full version with the six-tool price table, the comparison chart and the
Vietnam stack notes lives here:

https://tanit365.com/en/blog/posts/best-crm-for-ecommerce-2026/

Start from the failure mode, shortlist inside one category, and check the
billing unit before the feature list — that order saves more money than any
single discount.