When I was a kid, there was a small grocery shop on the corner of our apartment block. The shop assistant knew which bread my grandmother bought, knew my father only picked up the newspaper on Saturdays, and knew that whenever I had pocket money I would head straight for the wafer biscuits. She had no software. She had memory, attention and a real interest in seeing us come back. That was personalization, long before anyone gave it a name.
I have worked in online retail since 2005, from MarketOnline to Floria (an online flower shop), and since 2013 I have been building Retargeting with Rareș Bănescu, a marketing automation and personalization platform for online stores. Over these years I have seen plenty of shops claim they do personalization. Most of the time that means a “Hi John” at the top of the newsletter, while the rest of the email is identical to the one sent to tens of thousands of other people. That is not personalization. That is a database field dropped into the right place.
What personalization really means
Real personalization means every customer sees what is relevant to them, not what the store wants to push that week. The lady at the corner shop did not use my name to impress me. She put the wafers aside for me. The difference is between recognizing a customer and understanding them.
Amazon showed years ago what this looks like at scale: “customers who bought this item also bought”. Netflix rearranges your whole home screen after you watch two series. It is not magic and it is not only for the giants. It is the same logic as the shop assistant, applied to thousands of customers at once.
Recommendations based on browsing and purchase history
You already have the most valuable signals: what someone looked at, what they put in the basket, what they bought, what they returned. A visitor who opened the same pair of shoes five times is telling you something. A customer who bought a coffee machine three months ago probably needs filters or coffee. Someone who ordered a wedding dress does not want to see wedding dresses next week.
Good recommendations respect context. On a product page, show alternatives and accessories. In the basket, show complementary items, not ten more versions of the same thing that only make the customer hesitate. After the purchase, show what logically comes next. And, crucially, do not recommend the exact product the person just bought. It sounds obvious, yet I see ads doing exactly that every day.
A personalized homepage and search
Most stores have a static shop window for a homepage: the banner with this month’s promotion and a few products someone on the team picked. For a first-time visitor that is fine. For someone coming back for the fifth time, it is a missed opportunity. If you know this person always buys toys for toddlers, why show them power drills first?
Search works the same way. Two people typing “bag” may be looking for very different things. The order of results can reflect what each of them has viewed and bought before: brands, price ranges, sizes. You do not need to rebuild the whole site overnight. Start with a “recently viewed” block and a “recommended for you” block on the homepage. You will see the effect quickly in time spent on site.
Dynamic emails and timing
In my experience, email is still one of the most profitable channels an online store has. The problem is not the channel; it is that we send everyone the same message at the same hour. A truly personalized email is assembled at the moment of sending, for each recipient: the items left in the basket, products they viewed, a price drop on something they were watching, a missing size back in stock.
Timing matters as much as the message. An abandoned basket reminder goes out the same day, not a week later. Consumables can be reminded at the pace they get used up. With flowers, for instance, important dates repeat every year: someone who ordered for an anniversary usually appreciates a reminder a few days before the same date next year. Even the send time can follow the hour when each person usually opens their email.
Not everyone is equally price-sensitive
Some customers only buy on sale. Others have never used a coupon and buy new arrivals at full price. Send everyone a 20% code and you hand money back to people who would have bought anyway, while teaching the rest to wait for discounts. I have seen stores eat away at their own margins this way, year after year.
Personalization means this too: the discount goes to whoever needs a nudge, and the loyal customer who does not look at prices gets something else, such as early access to a new collection, free delivery or a simple thank-you. This is not about charging different prices for the same product, which rightly annoys people. It is about choosing the right incentive.
What data a small store already has
Many small store owners tell me personalization is for the big players with teams of analysts. It is not. A shop with a few hundred orders a month already knows far more than the corner shop assistant ever did:
- order history: what, how much, how often, at what value;
- products viewed and searches made on the site;
- abandoned baskets;
- email opens and clicks;
- where the customer came from: search, ads, social media, newsletter;
- delivery city, payment method and sometimes the date of a special occasion.
You do not need all of it from day one. Start with what you have and two or three simple scenarios: abandoned basket, post-purchase recommendations, winning back customers who have not bought in a while. Today there are tools that do this without hiring developers. We built Retargeting precisely because small stores had the data but no practical way to use it.
Respecting privacy is not optional
There is a fine line between “how nice of them to remind me” and “how on earth do they know that?”. The corner shop lady knew what I bought, but she did not follow me home. When customers feel watched, they stop feeling cared for and start feeling uneasy. A few rules I consider common sense: say clearly what data you collect and why, never use sensitive information for advertising, offer a real one-click unsubscribe, and never sell your customers’ data to anyone.
The law is heading the same way. In April this year the European Union adopted the new General Data Protection Regulation, which will apply from May 2018. The details will become clearer over the coming months, but the message is plain: clear consent, transparency and a person’s right to see and delete their data. Stores that get used to working cleanly now will have nothing to fear two years from now.
Personalization is an attitude, not a plug-in
Technology helps, but it does not replace the basic question: what would a good shopkeeper who knows their customers do in my place? They would remember what people like, not pester them with useless offers, tell them when something they were waiting for arrives, and leave them alone when they need nothing. If your software does that, you have personalization. If it only puts a name in the subject line, you have a template with ambitions.
What do you do in your store to make returning customers feel recognized? And as shoppers, what is the best, or the strangest, recommendation you have ever received online?