Segmentacija kupcev

Kako segmentirati kupce spletne trgovine glede na nakupno vedenje

Posted on July 19, 2026 by

Kupce segmentirajte glede na to, kaj v resnici počnejo, ne glede na to, kdo so. 35-letna ženska iz mesta in 42-letni moški s podeželja se lahko v vaši trgovini vedeta popolnoma enako — oba sta kupila enkrat, nobeden se ni vrnil. Starost in kraj bivanja vam o naslednji prodaji ne povesta skoraj nič. Njuno vedenje pa pove vse: kaj sta si ogledovala, kaj sta dodala v košarico, kaj sta kupila, iz katere kategorije, kako nedavno in kako pogosto. Ljudi razvrstite po teh dejanjih in vaša e-poštna sporočila nehajo biti ena in ista vsebina, zakričana vsem naenkrat. To je vedenjska segmentacija in za trgovino, ki že ustvarja stabilen prihodek, je to običajno najhitrejši način, da iztržite več iz seznama, ki ga že imate.

Zakaj demografija tiho troši vaš proračun

Večina trgovin začne z demografskimi segmenti, ker so jih tako naučile razmišljati oglaševalske platforme. Spol, starostna skupina, regija. Zdi se, kot da poznate svojo stranko.

Težava je, da to ne napoveduje nakupa. Dve osebi v isti starostni skupini si od vaše trgovine želita popolnoma različne stvari, nekdo, ki je pravkar zapustil košarico za 90 €, pa je bolj vroč kandidat kot popolno demografsko ujemanje, ki trgovine sploh še ni obiskalo. Ko celotnemu seznamu pošljete isti “tukaj je 10 % popusta”, plačujete — v marži in odjavah — za to, da dosežete ljudi, ki bi kupili tako ali tako, in jezite ljudi, ki bi potrebovali popolnoma drugačno sporočilo.

Vedenje to popravi, ker je vedenje namera, ki jo lahko vidite. Ogled izdelka je majhen glas. Dodajanje v košarico je večji. Zaključeno naročilo v določeni kategoriji vam pove, kaj si nekdo v resnici želi, z lastnimi besedami in z lastnim denarjem.

Vedenjski signali, ki jih je vredno upoštevati

Ne potrebujete ekipe podatkovnih znanstvenikov. Skoraj vsaka platforma za spletno prodajo in orodje za e-pošto že spremljata peščico signalov, ki so pomembni. Predstavljajte si jih kot osi, po katerih lahko režete — vsako zase ali v kombinaciji.

Angažiranost, še pred kakršnim koli nakupom. So odprli zadnjih pet e-poštnih sporočil? Kliknili karkoli? Obiskali stran ta mesec? Nekdo, ki redno brska, a ne kupuje, je drugačna težava kot nekdo, ki je utihnil. To si zasluži ločen pogled — glejte segmentacijo kupcev glede na stopnjo angažiranosti in segmentacijo glede na vedenje pri brskanju.

Nedavnost — koliko časa je minilo od zadnjega nakupa. Ta napoveduje osip bolje kot skoraj karkoli drugega. Kupec, ki je 20 dni od zadnjega nakupa, se vede popolnoma drugače kot tisti pri 200 dneh. Sama nedavnost lahko poganja cel program reaktivacije; mehaniko najdete v kako segmentirati kupce glede na čas od zadnjega nakupa.

Pogostost — kolikokrat so kupili. Eno naročilo v primerjavi s tremi je verjetno najpomembnejša črta v vsej vaši bazi. Loči neznance od stalnih strank in spremeni, kaj bi morali povedati vsakemu. Več v kako segmentirati kupce glede na pogostost nakupov.

Denarna vrednost — koliko so vredni. Skupna poraba, povprečna vrednost naročila, marža po vračilih. Vaši največji porabniki si zaslužijo drugačno obravnavo kot lovci na ugodnosti.

Kupljena kategorija in izdelek. Kar je nekdo kupil, je najčistejši signal, kaj mu pokazati naslednjič. Lastniki psov nočejo mačje hrane. To je hrbtenica ustreznih priporočil — segmentacija kupcev glede na kategorijo izdelkov to obravnava temeljito.

Prve tri — nedavnost, pogostost, denarna vrednost — se združijo v klasični model, ki ga sčasoma prevzame večina trgovin. Če želite strukturirano različico, je RFM-segmentacija standardni način, da kupce ocenite po vseh treh naenkrat.

Kje prihodek dejansko odteka

Tukaj je puščanje v jasnih številkah. Recimo, da pošljete eno tedensko kampanjo 8.000 stikom (za ponazoritev). Vsi dobijo isto kodo za 15 %. Vaši ponovni kupci in VIP-stranke — morda 900 od njih — bi ta mesec kupili po polni ceni tako ali tako. Ta popust je čista, zavržena marža. Medtem je 300 ljudi ta teden zapustilo košarice in dobilo isto splošno sporočilo namesto opomnika o točno tistem izdelku, ki so ga pustili. Nekaj tisoč zaspanih stikov, ki potrebujejo pravi razlog za vrnitev, pa je dobilo sporočilo, napisano za aktivne kupce.

Ista pošiljka. Tri različne napake. Nobena se ne pokaže kot napaka — kampanja je “delovala”, le da je v vseh smereh hkrati pustila denar na mizi.

Vedenjska segmentacija zamaši te luknje, ker vsaka skupina končno dobi sporočilo, ki ustreza točki, kjer se nahaja.

Kako začeti, ne da bi si zapletli življenje

Uprite se skušnjavi, da bi zgradili 40 segmentov. To je pogosta past in ustvarja vzdrževalno delo, ki mu nobena majhna ekipa ne more slediti. Začnite z delitvijo, ki se izplača najhitreje.

Ločite kupce od nekupcev. Vsak na vašem seznamu ima naročilo ali pa ga nima. Ti dve skupini potrebujeta nasprotni sporočili. Nekupci potrebujejo zaupanje in prvi razlog za nakup; kupci potrebujejo ustreznost in razlog za vrnitev.
Znotraj kupcev ločite enkratne od ponovnih. Ta ena črta preoblikuje celoten trud za zadržanje — segmentacija prvih in ponovnih kupcev poglobljeno pojasni, zakaj.
Označite skrajnosti. Vaše najvrednejše stranke na eni strani, najbolj ogrožene zaspane stranke na drugi. Obe sta majhni skupini z nesorazmerno velikim vplivom.
Ko je smiselno, dodajte še kategorijo. Ko delitev po življenjskem ciklu deluje, dodajte “kupil iz kategorije X”, da izostrite priporočila.

To je pet ali šest segmentov, ki opravljajo pravo delo, ne pa petdeset, ki nabirajo prah. Ko ste pripravljeni namensko zgraditi celoten začetni nabor, jih segmenti kupcev, ki bi jih morala ustvariti vsaka spletna trgovina predstavijo enega za drugim.

Kaj avtomatizirati najprej

Segmenti so uporabni šele, ko se sporočilo sproži iz njih, ne da bi vi kaj naredili. Najpomembnejša vedenja spremenite v stalne avtomatizacije:

Sprožilec: dodajanje v košarico, brez nakupa v eni uri. Segment: kateri koli stik. Zamik: 1 ura, nato 24 ur. Kanal: e-pošta (za dražje košarice dodajte SMS). Vsebina: točno tisti izdelek, spodbuda, pomoč, če so obtičali. Cilj: povrniti prodajo.
Sprožilec: zaključeno prvo naročilo. Segment: prvi kupci. Zamik: takojšnja zahvala, nato preverjanje čez teden dni. Cilj: približati jih drugemu naročilu.
Sprožilec: brez nakupa v X dneh (X nastavite na svojo tipično vrzel med ponovnimi naročili plus rezerva). Segment: zaspani kupci. Cilj: pridobiti jih nazaj, preden so za vedno izgubljeni.

Vedenje samodejno vstopa in izstopa iz teh segmentov. Kupite nekaj in samodejno izpadete iz toka za zapuščene košarice. To je bistvo grajenja na dejanjih namesto na statičnih seznamih.

Metrike, ki povedo, da deluje

Tega ne sodite po stopnji odpiranja. Sodite ga po denarju in zdravju:

Prihodek na prejemnika, primerjan med segmentirano pošiljko in množičnim razpošiljanjem. To je številka, ki razreši prepir.
Stopnja ponovnih nakupov — se tok od prvega do drugega naročila premika?
Povrnjene košarice iz avtomatizacije za zapuščene košarice.
Stopnja odjav in prijav neželene pošte — segmentacija bi ju morala potisniti navzdol, ker ljudje dobijo manj neustreznih sporočil.

Če želite čisto metodo za primerjavo prej in potem, kako izmeriti, ali segmentacija izboljša prihodek pravilno postavi primerjavo.

Kam se umešča Omnisend

Ko se odločite, katera vedenja upoštevati, potrebujete orodje, ki jih spremlja in samodejno sproži pravi tok. Preizkusil sem tako Klaviyo kot Omnisend, preden sem se za svoje trgovine odločil za Omnisend — predvsem zaradi tega, kako hitro sem lahko sam zgradil vedenjske segmente, brez specialista, in v istem toku združil e-pošto, SMS in potisna sporočila. Segmenti se tam posodabljajo v živo: stik izpolni pogoje in se sam premakne noter ali ven, kar je točno to, kar vedenjska segmentacija potrebuje, da ostane natančna.

Poštena omejitev: nič od tega ne reši slabe ponudbe ali tankega seznama. Če izdelek ne ustreza trgu, boljši segmenti le natančneje dostavijo povprečno sporočilo. Avtomatizacija izostri tisto, kar že deluje; povpraševanja ne ustvari. In sledenje je le tako dobro, kot so podatki, ki pritekajo, zato je čista integracija med trgovino in e-pošto pomembnejša od katere koli posamezne funkcije.

Razkritje o partnerstvu: nekatere povezave do Omnisend so partnerske povezave. Če se prijavite prek njih, lahko Shopimation prejme provizijo brez dodatnih stroškov za vas. Priporočam le orodja, ki jih dejansko uporabljam.

Ustrezno, na vedenju temelječe sporočanje je tudi temelj za personalizacijo in priporočila izdelkov — segmentacija odloči, kdo dobi kaj, personalizacija pa odloči, katere izdelke vidijo.

Vaš naslednji korak

Ta teden potegnite eno številko: koliko stikov na vašem seznamu ima nič naročil, eno naročilo in dve ali več naročil. Ta en izračun je vaš prvi vedenjski zemljevid in običajno natančno pokaže, kam bi morala iti vaša naslednja kampanja. Iz njega zgradite en segment — začnite s prvimi kupci — in jim pošljite nekaj, česar preostali seznam ne dobi.

Customer Segmentation

Segmenting Ecommerce Customers by Purchase Behavior

Posted on July 17, 2026July 17, 2026 by

Segment your customers by what they actually do, not who they are. A woman aged 35 in a city, and a man aged 42 in the countryside, might behave identically in your store — both bought once, both never came back. Age and location tell you almost nothing about your next sale. Their behavior tells you everything: what they viewed, what they added to cart, what they bought, which category they bought it in, how recently, and how often. Group people by those actions and your emails stop being one message shouted at everyone. That’s behavioral segmentation, and for a store already doing steady revenue it’s usually the fastest way to earn more from the list you already have.

Why demographics quietly waste your budget

Most stores start with demographic segments because that’s what ad platforms trained them to think about. Gender, age bracket, region. It feels like knowing your customer.

The problem is that it doesn’t predict buying. Two people in the same age band want completely different things from your store, and someone who just abandoned a €90 cart is a hotter prospect than a perfect demographic match who has never visited. When you send the same “here’s 10% off” blast to your whole list, you’re paying — in margin and in unsubscribes — to reach people who were going to buy anyway, and annoying people who needed a different message entirely.

Behavior fixes this because behavior is intent made visible. A product view is a small vote. An add-to-cart is a bigger one. A completed order in a specific category tells you what someone actually wants, in their own words, with their own money.

The behavioral signals worth acting on

You don’t need a data science team. Almost every ecommerce platform and email tool already tracks the handful of signals that matter. Think of them as axes you can slice along, alone or in combination.

Engagement, before any purchase. Did they open the last five emails? Click anything? Visit the site this month? Someone browsing regularly but not buying is a different problem than someone who went silent. This is worth a closer look on its own — see segmenting customers by engagement level and segmenting by browsing behavior.

Recency — how long since the last purchase. This one predicts churn better than almost anything. A customer 20 days out behaves nothing like one at 200 days. Recency alone can drive a whole reactivation program; the mechanics live in how to segment customers by time since last purchase.

Frequency — how many times they’ve bought. One order versus three orders is arguably the most important line in your whole database. It splits strangers from regulars and changes what you should say to each. More in how to segment customers by purchase frequency.

Monetary value — what they’re worth. Total spend, average order value, margin after returns. Your top spenders deserve different treatment than your bargain hunters.

Category and product bought. What someone purchased is the cleanest signal of what to show them next. Dog owners don’t want cat food. This is the backbone of relevant recommendations — segmenting customers by product category covers it properly.

The first three — recency, frequency, monetary — combine into the classic model most stores eventually adopt. If you want the structured version, RFM segmentation is the standard way to score customers on all three at once.

Where the revenue actually leaks

Here’s the leak in plain numbers. Say you send one weekly campaign to 8,000 contacts (illustrative). Everyone gets the same 15% code. Your repeat buyers and VIPs — maybe 900 of them — would have bought at full price this month anyway. That discount is pure margin thrown away. Meanwhile 300 people abandoned carts this week and got the same generic email instead of a reminder about the exact product they left behind. And a few thousand lapsed contacts, who need a real reason to return, got a message written for active shoppers.

Same send. Three different failures. None of them show up as an error — the campaign “worked,” it just left money on the table in every direction at once.

Behavioral segmentation plugs those leaks because each group finally gets the message that fits where they are.

How to start without overcomplicating it

Resist the urge to build 40 segments. That’s a common trap, and it creates maintenance work no small team can keep up with. Start with the split that pays off fastest.

Split buyers from non-buyers. Everyone on your list either has an order or doesn’t. These two groups need opposite messages. Non-buyers need trust and a first reason to buy; buyers need relevance and a reason to return.
Inside buyers, split one-time from repeat. This single line reshapes your whole retention effort — segmenting first-time and repeat buyers goes deep on why.
Flag the extremes. Your highest-value customers on one end, your most at-risk lapsing customers on the other. Both are small groups with outsized impact.
Layer category on top when it’s relevant. Once the lifecycle split works, add “bought from X category” to make recommendations sharper.

That’s five or six segments doing real work, not fifty gathering dust. When you’re ready to build the full starter set deliberately, the customer segments every online store should create lays them out one by one.

What to automate first

Segments are only useful when a message fires off them without you lifting a finger. Turn the top behaviors into standing automations:

Trigger: add-to-cart, no purchase within an hour. Segment: any contact. Delay: 1 hour, then 24 hours. Channel: email (add SMS for higher-value carts). Content: the exact product, a nudge, help if they got stuck. Goal: recover the sale.
Trigger: first order completed. Segment: first-time buyers. Delay: immediate thank-you, then a check-in a week later. Goal: move them toward a second order.
Trigger: no purchase in X days (set X to your typical reorder gap plus a margin). Segment: lapsing buyers. Goal: win them back before they’re gone for good.

Behavior enters and leaves these segments on its own. Buy something, and you drop out of the abandoned-cart flow automatically. That’s the point of building on actions instead of static lists.

Metrics that tell you it’s working

Don’t judge this by open rate. Judge it by money and health:

Revenue per recipient, compared between a segmented send and a batch-and-blast. This is the number that settles the argument.
Repeat purchase rate — is the first-to-second-order flow moving?
Recovered carts from the abandonment automation.
Unsubscribe and spam rate — segmentation should push these down, because people get fewer irrelevant emails.

If you want a clean before-and-after method, how to measure whether segmentation improves revenue sets up the comparison properly.

Where Omnisend fits

Once you’ve decided which behaviors to act on, you need a tool that tracks them and fires the right flow automatically. I tested both Klaviyo and Omnisend before settling on Omnisend for my own stores — mostly for how quickly I could build behavioral segments myself, without a specialist, and combine email, SMS, and push in the same flow. Segments there update live: a contact meets the conditions and moves in or out on their own, which is exactly what behavioral segmentation needs to stay accurate.

Honest limit: none of this rescues a weak offer or a thin list. If your product-market fit isn’t there, better segments just deliver a mediocre message more precisely. Automation sharpens what already works; it doesn’t manufacture demand. And the tracking is only as good as the data flowing in, so clean store-to-email integration matters more than any single feature.

Affiliate disclosure: some links to Omnisend are affiliate links. If you sign up through them, Shopimation may earn a commission at no extra cost to you. I only recommend tools I actually use.

Relevant, behavior-based messaging is also the foundation for personalization and product recommendations — segmentation decides who gets what, personalization decides which products they see.

Your next step

Pull one number this week: how many contacts on your list have zero orders, one order, and two-plus orders. That single count is your first behavioral map, and it usually shows exactly where your next campaign should go. Build one segment from it — start with first-time buyers — and send them something the rest of your list doesn’t get.