Od prvega do drugega nakupa

Kako meriti čas med prvim in drugim nakupom

Posted on July 18, 2026 by

Najčistejši način, da izmerite čas med prvim in drugim nakupom kupca, je, da vzamete vsakega kupca, ki je naročil vsaj dvakrat, izračunate število dni med datumom prvega in datumom drugega naročila ter nato pogledate mediano teh razmikov – ne povprečja. Mediana vam pove dan, do katerega se je vrnila polovica vaših ponovnih kupcev. Prav ta ena številka je tista, okoli katere gradite časovnico po nakupu. Če je vaša mediana razmika 34 dni, je ponudba, ki pristane na 3. dan, prezgodnja, ponudba na 90. dan pa prepozna. Ta članek govori zgolj o tem, kako to številko pravilno izračunati in jo pošteno prebrati. Kaj z njo nato storite – torej sam čas ponudbe za drugi nakup – je ločena odločitev, na katero vas napotim na koncu.

Številka, ki jo v resnici želite, na jasno opredeljena

Obstajajo tri sorodne številke, ki jih ljudje mešajo med seboj, zato jih ločimo.

Čas do drugega nakupa – dnevi med naročilom 1 in naročilom 2, merjeno samo pri kupcih, ki so prišli do naročila 2. To je tista, o kateri govori ta članek.
Stopnja prehoda od prvega do drugega nakupa – delež prvih kupcev, ki kdaj oddajo drugo naročilo. Druga metrika, obravnavana v kako izgleda zdrava stopnja prehoda od prvega do drugega nakupa.
Povprečni razmik med naročili – razmik med vsemi ponovnimi naročili, od 2 do 3, od 3 do 4 in tako naprej. Uporabno za dopolnjevanje zalog, a zabriše prvi razmik, ki se obnaša drugače kot vsak razmik za njim.

Prvo ohranite čisto. Štejejo samo kupci z dvema ali več naročili, saj prvi kupec, ki se še ni vrnil, nima razmika, ki bi ga lahko izmerili – če ga vključite kot ničlo ali ogromno število, uničite številko.

Zakaj povprečje laže, mediana pa ne

Recimo, da izvlečete razmik za sto ponovnih kupcev. Večina se vrne nekje med tremi in osmimi tedni. A peščica jih je znova naročila enajst mesecev pozneje – pozabili so na vas, nato pa jih je nazaj pripeljal opomnik za rojstni dan ali nenadna potreba.

Teh nekaj dolgih repov potegne povprečje močno navzgor. Lahko se znajdete s »povprečnim časom do drugega nakupa« 70 dni, čeprav se je tri četrtine vaših ponovnih kupcev dejansko vrnilo v manj kot 40 dneh. Če časovnico e-poštnih sporočil gradite na 70, ljudem pošiljate sporočila dva tedna po tem, ko bi že kupili.

Mediana ne upošteva osamelcev. Je srednja vrednost: postavite vse razmike v vrsto od najkrajšega do najdaljšega in vzemite tistega na sredini. Za časovnico po nakupu je mediana pošteno sidro. Pogledal bi tudi 25. in 75. percentil – dneve, do katerih se je vrnila četrtina in tri četrtine ponovnih kupcev – saj vam ta razpon pove, kako tesno ali ohlapno je vaše okno za ponovno naročilo.

Kje se skriva izguba, ko to preskočite

Večina trgovin razmika sploh nikoli ne izmeri. Čas pošiljanja po nakupu izberejo po občutku – »pošljimo opomnik po enem tednu« – in ga nikoli ne preverijo glede na to, kako se njihovi kupci dejansko obnašajo.

Tu je tihi strošek. Predstavljajte si trgovino, ki prodaja premium pasjo hrano. Vreča za srednje velikega psa zdrži približno pet tednov. Če ta trgovina pošlje e-pošto »pripravljeni za novo naročilo?« na 7. dan, se pogovarja s kupcem, ki ima omaro še vedno polno – sporočilo se prezre, ko pa vreča na 33. dan poide, ne pride noben opomnik. Ponovno naročilo, ki bi moralo biti samodejno, se zavleče, in nekateri od teh kupcev zaidejo k tistemu oglasu, ki jih prvi ujame, ko jim zmanjka.

Razmik med »kdaj pošljemo e-pošto« in »kdaj so pripravljeni« je čist odtekli prihodek. Ne morete ga zapreti, dokler ne izmerite, kam v resnici pade trenutek pripravljenosti. In ta se ogromno razlikuje glede na to, kaj prodajate – naročnik na kavo in kupec vzmetnice živita po povsem različnih urah, zato je ob tem vredno prebrati tudi kako vrsta izdelka spremeni čas ponudbe za drugi nakup.

Kako to izračunati, korak za korakom

Ne potrebujete podatkovne ekipe. Potrebujete izvoz.

Izvozite naročila z vsaj ID-jem kupca (ali e-pošto), datumom naročila in zaporedjem naročila. Večina platform omogoča izvoz vseh naročil kot CSV.
Obdržite samo kupce z dvema ali več naročili. Razvrstite po kupcu, nato po datumu. Vsak z enim samim naročilom gre v stran za metriko stopnje, ne za to.
Za vsakega kupca poiščite datuma naročila 1 in naročila 2. Prvi datum odštejte od drugega. To je njihov razmik v dneh.
Naštejte vse te razmike v en stolpec in vzemite mediano. V preglednici je to =MEDIAN(obseg). Ko ste že pri tem, vzemite še =PERCENTILE(obseg,0.25) in =PERCENTILE(obseg,0.75) za razpon.
Preverite skrajnosti. Razmik 0 dni običajno pomeni dve naročili istega dne (pogosto popravek ali razdeljeno pošiljko) – odločite se, ali ju združite ali izpustite. Razmik 400+ dni je pristen povratnik; ohranite ga v podatkih, a vedite, da vas mediana pred njim že varuje.

To je celotno delo. Petnajst minut z enim izvozom.

Berite po kohortah, ne kot eno gmoto

Ena mediana za celotno trgovino je dober začetek, a uporaben vpogled pride iz razrezovanja.

Razčlenite razmik po prvem izdelku ali prvi kategoriji. V mnogih trgovinah vstopni izdelek tiho napoveduje uro ponovnega naročila: nekdo, čigar prvi nakup je bil potrošno blago, se vrača v tesnem, predvidljivem ciklu, medtem ko nekdo, čigar prvi nakup je bila enkratna naprava, morda potrebuje mesece ali se v tej kategoriji nikoli ne vrne. Če vaši podatki kažejo tak razkorak, ste našli nekaj, na kar lahko ukrepate – različni prvi izdelki si zaslužijo različno časovnico nadaljnjih sporočil.

Spremljajte tudi razmik skozi čas. Izvlecite ga za kupce, ki so prvič kupili to četrtletje, v primerjavi s tistimi izpred leta. Če se razmik pri nedavni kohorti razteza, se je nekaj spremenilo – dvig cen, težava z zalogo, šibkejše nadaljnje sporočanje po nakupu – in številka vas zgodaj opozarja.

Preprost izdelan primer (ponazoritveni)

Recimo, da izvozite in ugotovite te razmike med prvim in drugim nakupom, v dneh, za deset ponovnih kupcev: 12, 19, 21, 26, 28, 31, 33, 40, 58, 240.

Povprečje je 50,8 dni – potegnjeno navzgor skoraj v celoti zaradi tistega osamljenega 240.
Mediana je 29,5 dni – sredina med 5. in 6. vrednostjo.

Devetindvajset dni je resnica, okoli katere bi načrtovali; enainpetdeset je fatamorgana. To je izmišljen nabor, ki prikazuje mehaniko – poženite ga na svojih naročilih in oblika bo drugačna, a nauk drži: pustite, da vodi mediana. (Ponazoritvene številke.)

Katere metrike spremljati poleg nje

Razmik je sam po sebi vhodni podatek za časovnico. Povežite ga z dvema ali tremi drugimi, da ostane pošten:

Mediana razmika med prvim in drugim nakupom – vaše sidro za časovnico ponovnega naročila, osveženo četrtletno.
Razpon med 25. in 75. percentilom – kako široko je vaše okno za ponovno naročilo; tesen razpon pomeni ostro okno pošiljanja, širok pa daljše negovanje.
Stopnja prehoda od prvega do drugega nakupa – ali se ljudje sploh vračajo, kar nobena rešitev časovnice sama ne more rešiti.
Razmik po prvem izdelku – kje se mora vaša časovnica nadaljnjih sporočil razlikovati.

Če je razmik videti v redu, stopnja pa pada, čas ni vaša težava – zadržanje je, in to je diagnoza v kaj spremeniti, ko stopnja ponovnega nakupa po prvem naročilu pade.

Kako Omnisend to naredi merljivo in uporabno

Mediano lahko dobite iz preglednice, in tam bi tudi začel – ne stane nič in vas prisili, da pogledate svoje surove podatke. Kjer si orodje zasluži svoje mesto, je pretvarjanje te številke v časovnico, ki se sproži sama.

V Omnisendu so podatki o življenjskem ciklu in nakupih kupca na profilu, tako da lahko, ko poznate svojo mediano razmika, nastavite avtomatizacijo po nakupu, da počaka pravo število dni pred naslednjim sporočilom, ne pa ugibanega. Njegova segmentacija vam omogoča tudi delitev kupcev po prvem izdelku ali kategoriji, kar je natanko tisti razrez, ki naredi razmik uporaben za ukrepanje. V svojih trgovinah ga uporabljam za to, potem ko sem ga preizkusil proti Klaviyu; razlog, da sem ostal, je bila vsakodnevna samopostrežnost – gradnja segmenta in prilagajanje zamika nista potrebovala strokovnjaka. Omnisend je partner Shopimationa v pridruženem programu in ga priporočam iz vsakodnevne uporabe. Vendar razmika ne bo izmeril namesto vas kar sam od sebe; številko še vedno najprej izvlečete sami, nato pa pustite orodju, da na njeni podlagi ukrepa.

Vaš naslednji korak

Ta teden izvozite svoja naročila in izračunajte eno številko: mediano dni med prvim in drugim nakupom. Zapišite jo na lepljiv listek. Nato to številko odnesite v odločitev o časovnici – kako kmalu naj prvega kupca prosite za ponoven nakup – in ko boste pripravljeni, da se sproži samodejno, v avtomatizacijo drugega nakupa, ki bi jo morala zgraditi vsaka rastoča trgovina.

First-to-Second Purchase

How to Measure the Time Between First and Second Purchases

Posted on July 18, 2026 by

The cleanest way to measure the time between a customer’s first and second purchase is to take every customer who has ordered at least twice, calculate the number of days between their first order date and their second order date, and then look at the median of those gaps — not the average. The median tells you the day by which half your repeat buyers have come back. That single number is what you build your post-purchase timing around. If your median gap is 34 days, an offer that lands on day 3 is too early and an offer on day 90 is too late. This article is only about how to calculate that number correctly and read it honestly. What you then do with it — the timing of the second-purchase offer itself — is a separate decision I’ll point you to at the end.

The number you actually want, defined

There are three related figures people mix up, so let’s separate them.

Time to second purchase — days between order 1 and order 2, measured only across customers who reached order 2. This is the one this article is about.
First-to-second purchase rate — the share of first-time buyers who ever place a second order. Different metric, covered in what a healthy first-to-second purchase rate looks like.
Average order gap — the spacing across all repeat orders, order 2 to 3, 3 to 4, and so on. Useful for replenishment, but it blurs the first gap, which behaves differently from every gap after it.

Keep the first one clean. Only customers with two or more orders count, because a first-time buyer who hasn’t returned yet has no gap to measure — including them as a zero or a giant number wrecks the figure.

Why the average lies and the median doesn’t

Say you pull the gap for a hundred repeat customers. Most come back somewhere between three and eight weeks. But a handful ordered again eleven months later — they’d forgotten you, then a birthday reminder or a random need brought them back.

Those few long tails drag the average way up. You could end up with an “average time to second purchase” of 70 days when three-quarters of your repeat buyers actually returned inside 40. Build your email timing on 70 and you’re mailing people two weeks after they’d already have bought.

The median ignores the outliers. It’s the middle value: line every gap up from shortest to longest and take the one in the center. For post-purchase timing, the median is the honest anchor. I’d also glance at the 25th and 75th percentile — the days by which a quarter and three-quarters of repeat buyers have returned — because that spread tells you how tight or loose your reorder window is.

Where the loss hides when you skip this

Most stores never measure the gap at all. They pick a post-purchase send time by gut — “let’s follow up after a week” — and never check it against how their customers actually behave.

Here’s the quiet cost. Imagine a store selling premium dog food. The bag lasts about five weeks for a medium dog. If that store sends its “ready for another order?” email on day 7, it’s talking to a customer whose cupboard is still full — the message gets ignored, and by the time the bag runs low on day 33, no reminder arrives. The reorder that should have been automatic drifts, and some of those customers wander to whatever ad catches them first when they run out.

The gap between “when we email” and “when they’re ready” is pure leaked revenue. You can’t close it until you’ve measured where the ready moment actually falls. And it varies enormously by what you sell — a coffee subscriber and a mattress buyer live on completely different clocks, which is why how product type changes the timing of the second-purchase offer is worth reading alongside this.

How to calculate it, step by step

You don’t need a data team. You need an export.

Pull an orders export with, at minimum, customer ID (or email), order date, and order sequence. Most platforms let you export all orders as a CSV.
Keep only customers with two or more orders. Sort by customer, then by date. Everyone with a single order gets set aside for the rate metric, not this one.
For each customer, find order 1 and order 2 dates. Subtract the first date from the second. That’s their gap in days.
List all those gaps in one column and take the median. In a spreadsheet that’s =MEDIAN(range). While you’re there, grab =PERCENTILE(range,0.25) and =PERCENTILE(range,0.75) for the spread.
Sanity-check the extremes. A gap of 0 days usually means two orders on the same day (often a correction or a split shipment) — decide whether to merge or drop those. A gap of 400+ days is a genuine returner; keep it in the data but know the median already protects you from it.

That’s the whole job. Fifteen minutes with one export.

Read it by cohort, not as one blob

One store-wide median is a fine start, but the useful insight comes from cutting it.

Break the gap down by first product or first category. In a lot of stores the entry product quietly predicts the reorder clock: someone whose first buy was a consumable comes back on a tight, predictable cycle, while someone whose first buy was a one-off gadget may take months or never return in that category. If your data shows that split, you’ve found something you can act on — different first products deserve different follow-up timing.

Also watch the gap over time. Pull it for customers who first bought this quarter versus a year ago. If the recent cohort’s gap is stretching, something changed — a price rise, a stock problem, a weaker post-purchase follow-up — and the number is warning you early.

A simple worked example (illustrative)

Say you export and find these first-to-second gaps, in days, for ten repeat customers: 12, 19, 21, 26, 28, 31, 33, 40, 58, 240.

The average is 50.8 days — dragged up almost entirely by that lone 240.
The median is 29.5 days — the midpoint between the 5th and 6th values.

Twenty-nine days is the truth you’d plan around; fifty-one is a mirage. This is a made-up set to show the mechanic — run it on your own orders and the shape will differ, but the lesson holds: let the median lead. (Illustrative figures.)

Which metrics to track alongside it

The gap on its own is a timing input. Pair it with two or three others so it stays honest:

Median first-to-second gap — your reorder-timing anchor, refreshed quarterly.
25th/75th percentile spread — how wide your reorder window is; a tight spread means a sharp send window, a wide one means a longer nurture.
First-to-second purchase rate — whether people are returning at all, which no timing fix can rescue on its own.
Gap by first product — where your follow-up timing needs to differ.

If the gap looks fine but the rate is falling, timing isn’t your problem — retention is, and that’s the diagnosis in what to change when repeat purchase rate drops after the first order.

How Omnisend makes this measurable and usable

You can get the median from a spreadsheet, and I’d start there — it costs nothing and forces you to look at your own raw data. Where a tool earns its place is turning that number into timing that fires on its own.

In Omnisend, customer lifecycle and purchase data sit on the profile, so once you know your median gap you can set a post-purchase automation to wait the right number of days before the next message rather than a guessed one. Its segmentation also lets you split customers by first product or category, which is exactly the cut that makes the gap actionable. I use it in my own stores for this after testing it against Klaviyo; the reason I stayed was the everyday self-service — building a segment and adjusting a delay didn’t need a specialist. Omnisend is an affiliate partner of Shopimation, and I recommend it from daily use. It won’t measure the gap for you out of the box, though; you still pull the number first, then let the tool act on it.

Your next step

Export your orders this week and calculate one figure: the median days between first and second purchase. Write it on a sticky note. Then take that number into the timing decision — how soon should you ask a first-time buyer to purchase again — and, when you’re ready to make it fire automatically, into a second-purchase automation every growing store should build.