Adding a customer's first name to a subject line is not personalisation. It is a mail merge, and shoppers stopped being impressed by it years ago. Real personalisation in ecommerce email means showing a specific customer the specific products they are most likely to want, at the moment they are most likely to want them. That is a data problem and a timing problem, and it is worth far more than a merge tag.
This article explains what personalised product recommendations actually are, what data drives them, which types work where, and how to tell whether any of it is making money. Two things have changed enough since this topic was last worth writing about that the old playbook now misleads more than it helps, and both are covered below.
What personalised product recommendations are
A personalised product recommendation is a product shown to a customer because something about that customer suggests they want it. The suggestion comes from data: what they bought, what they browsed, what they left in a cart, what people like them tend to buy next. Inside an email, this usually appears as a block of products chosen for that recipient rather than picked by hand for everyone.
There is a useful split here. Cosmetic personalisation changes the wrapping: a name in the subject, a location in the greeting, a "we thought of you" that means nothing. Substantive personalisation changes the contents: the actual products differ from one recipient to the next because their behaviour differs. Only the second kind moves revenue, and it is the only kind worth the effort of building.
The data that makes recommendations work
Good recommendations are only as good as the signals behind them. Five sources do most of the work.
Purchase history tells you what someone has already committed money to, which is the strongest signal you have. It supports replenishment, complementary products and sensible next purchases.
Browse behaviour tells you what someone considered but did not buy. A shopper who viewed three pairs of boots and left has told you what they want without spending anything.
Cart contents tell you about intent that stalled. Something stopped the purchase, and a recommendation paired with a reminder can restart it.
Category and brand affinity tells you the aisles a customer lives in. A person who only ever buys menswear should never receive a womenswear recommendation, and getting this wrong signals that you are not paying attention.
Behaviour of similar customers fills the gaps for people you do not yet know well. If shoppers who bought this product frequently went on to buy that one, that pattern is a reasonable guess for the next person who buys the first.
The email address is what ties these together, because it links the order, the browse session and the profile into one view of a person. Without that link, you are personalising to fragments.
Types of product recommendation and where each fits
Different recommendation logic suits different moments. Matching the type to the context is most of the skill.
Recently viewed and browse-based. Show people what they looked at and did not buy. This is among the highest-converting recommendation types because it reflects stated interest, not a guess. It belongs in browse-abandonment emails and in the top of a re-engagement message.
Complementary and cross-sell. Recommend what goes with something already bought or in the cart: the belt for the trousers, the case for the phone. This works best after purchase and inside the cart, where the base item makes the addition obvious.
Replenishment. For consumables, recommend the reorder timed to when the customer is likely running low. A skincare buyer who purchased a 60-day supply should hear from you around day 50, not day 5.
Upgrade and next-tier. Suggest the better or newer version to a customer whose history shows they buy up. Aim this carefully, since it lands badly on someone who only ever buys entry-level or on sale.
Catalogue-driven. Trigger a recommendation off a product event the customer cares about: back in stock in their size, a price drop on something they viewed, new arrivals in a category they favour. These convert well because the email exists for a reason the customer already had.
Trending and bestsellers. When you know little about a recipient, fall back to what is popular. It is a weak signal compared with behaviour, but it beats a random selection and covers new subscribers you have no history on.
Campaigns versus flows: where recommendations earn their money
Personalised recommendations pay off far more inside automated flows than inside one-off campaigns, and understanding why reshapes where you spend effort.
A campaign is a batch send: one email, many recipients, sent on a schedule. A flow is a triggered email: it fires when a specific customer does a specific thing, so it reaches them at a moment that already means something. Industry benchmark data consistently shows triggered flows generating far more revenue per recipient than batch campaigns, despite making up a small share of total send volume, because relevance and timing compound. A cart-abandonment email carrying the abandoned product and a complementary suggestion arrives while the intent is still warm. A birthday blast to your whole list does not.
The flows where recommendations do the most work are the predictable ones: welcome, browse abandonment, cart abandonment, post-purchase, replenishment and win-back. Build the recommendation logic into these first. A well-built flow keeps earning quietly for months without another hour of your time, which is the opposite of the campaign treadmill.
ing Scalable Systems
One of the most important aspects of scaling a company is creating systems that can handle growth. Many businesses rely on informal processes during their early stages, but as the company expands, those processes can become unreliable. Strategic advisors help businesses design structured systems for operations, communication, and decision-making. These systems allow companies to grow without losing efficiency or control.
Why Scaling a Business Is Challenging
Many founders start their companies with a strong vision, but growth introduces complexity. Processes that worked for a small team may begin to break down as operations expand.
- Unclear operational structures
- Inefficient workflows
- Compliance and regulatory issues
- Poor financial planning
- Limited strategic direction
Without the right systems in place, businesses often experience slower growth and operational bottlenecks.
Improving Financial Strategy
Financial planning is a key factor in sustainable growth. Advisors help businesses develop clear financial strategies that align with their expansion goals. This can include budgeting, cost management, tax planning, and investment decisions. By maintaining strong financial discipline, businesses can avoid common growth-stage problems such as overspending or poor resource allocation.
Why open rate is no longer how you measure this
Open rate used to be the headline number for email, and it is now the least trustworthy metric on the dashboard. The reason is Apple's Mail Privacy Protection, live since September 2021.
Mail Privacy Protection pre-loads email content, including the invisible tracking pixel that records an open, whether or not the recipient ever looks at the message. Apple Mail accounts for around half of all tracked opens, and a large majority of consumer subscribers use a Mail Privacy Protection capable client, so a big share of your reported opens are machines, not people. Open rates in Apple-heavy audiences are inflated by double digits, and in some segments most reported opens never happened.
This breaks more than a vanity number. Anything built on opens now misfires: subject-line A/B tests judged on open rate, "resend to non-openers" logic, and segmenting your "most engaged" by who opens. It also means the old advice to stuff a name in the subject line to lift opens is chasing a metric that no longer measures anything real.
Measure what the recipient actually did instead. Clicks, conversions and revenue per recipient survive the privacy changes because they record real actions. Revenue per recipient is the number that matters most for recommendations, because it tells you whether the products you showed turned into money.
Deliverability is the part nobody mentions
None of this matters if the email lands in spam, and deliverability is where a surprising number of "email isn't working" problems actually begin. Since February 2024, Google and Yahoo have enforced a shared set of requirements on bulk senders, roughly those sending 5,000 or more messages a day to their users, and meeting them is now the price of reaching the inbox.
The requirements are specific. Authenticate your mail with SPF and DKIM, publish a DMARC record, and make sure your sending domain aligns. Include a working one-click unsubscribe in marketing emails and honour it within two days. Keep your spam complaint rate below 0.3 percent, with under 0.1 percent the real target. Microsoft has since aligned with the same thresholds. Fail these and mail gets throttled, junked or rejected outright, at which point your beautifully personalised recommendations reach nobody.
Most ecommerce email platforms handle the authentication headers for you, but the responsibility for a clean list and a low complaint rate is yours. The fastest way to wreck deliverability is to email people who did not ask to hear from you, because the spam complaints that follow teach the inbox providers to distrust you. Permission and relevance are deliverability features, not just courtesies.
Personalisation that helps versus personalisation that unsettles
Recommendations build trust when they feel helpful and erode it when they feel like surveillance. The line sits at what the customer expects you to know.
Recommending a size you have on file, a reorder of something they buy regularly, or a product that complements their last purchase reads as attentive, because the customer understands how you knew. Referencing a single product they glanced at once, or personalising in a way that exposes how closely you have been tracking them, reads as watching. The discomfort is not about accuracy. It is about a shopper suddenly seeing the machinery.
The safe rule is to personalise from what a customer has clearly given you or plainly signalled, in service of making their next purchase easier, and to stop short of showing off how much you have observed. Useful beats clever every time.
Common mistakes to avoid
- Treating a name in the subject line as personalisation. It is cosmetic, it is expected, and it does nothing your competitors are not also doing. Personalise the products, not the greeting.
- Judging performance by open rate. Since Apple's privacy changes, open rate is close to noise. Measure clicks, conversions and revenue per recipient.
- Recommending the thing they just bought. Nothing says "we are not really paying attention" like suggesting the product sitting in the customer's cupboard. Suppress recent purchases from recommendations.
- Batch-blasting identical recommendations to everyone. If every recipient sees the same products, you have built a newsletter, not a recommendation. The value is in the difference between recipients.
- Ignoring deliverability. The best recommendation in the world earns nothing from the spam folder. Authenticate properly and protect your complaint rate.
- Personalising into surveillance. Relevance from data the customer gave you builds trust. Cleverness that reveals how closely you have watched them spends it.
How to tell whether it is working
Judge personalised recommendations by revenue, and by where that revenue comes from.
Revenue per recipient is the cleanest measure, because it folds deliverability, relevance and conversion into one honest number and it survives the open-rate problem. Track it separately for flows and campaigns, since a healthy flow figure can hide weak campaigns and the reverse. The share of your total email revenue that comes from automated flows tells you whether your triggered, personalised messages are pulling their weight, and in a well-run ecommerce program that share is substantial. Where your platform allows it, look at the revenue attributed to the recommendation blocks themselves, so you know the products you chose, rather than the rest of the email, are what converted.
Then watch click-to-conversion on the emails carrying recommendations. A high click rate with weak conversion usually means the email promised something the landing experience did not deliver, which is a merchandising or product-page problem rather than an email one.
Start with one flow, most usefully cart abandonment or post-purchase, and build genuine product recommendations into it before touching anything else. Get that earning, measure it by revenue per recipient, and move to the next flow. Across the ecommerce brands Growth Huntr works with, the pattern is consistent: the money is in a handful of well-built flows with real recommendation logic, not in a busier campaign calendar. If you want a clear read on which of your flows are underperforming and why before you rebuild anything, that is a conversation we are glad to have with you.
.svg)
.svg)