Povečanje povprečne vrednosti naročila

Kaj priporočiti, ko kupec v košarico doda prvi izdelek

Posted on July 18, 2026 by

V trenutku, ko nekdo v košarico doda prvi izdelek, priporoči izdelek, zaradi katerega ta izdelek deluje – dodatek, ki ga potrebuje, potrošni material, na katerem teče, ali del, ki dokonča delo, po katerega je prišel. Ne uspešnice. Ne večje različice iste stvari. Ne šestih naključnih izdelkov. Eno ali dve stvari, ki bi mu ju za pultom podal dobro poučen prodajalec, ker bi odhod brez njiju pomenil vrnitev nejevoljnega kupca. Ta prvi dodatek v košarico je signal namere, ki ga tako čisto le redko dobiš, in kar pokažeš zatem, deluje bodisi kot pomoč bodisi kot prodajni nagovor. Ta stran govori natanko o tem, kaj postaviti v to mesto – kateri izdelki, v kakšnem vrstnem redu in čemu se izogniti.

Zakaj je ta trenutek drugačen od vsakega drugega priporočila

Prvi dodatek v košarico ti pove tri stvari hkrati: kaj kupec želi, da je pripravljen kupiti in da je še vedno v nakupovalni miselnosti, ne v plačilni. Ta kombinacija je redka. Na domači strani ugibaš. Pri zaključku nakupa je odločitev v mislih že zaprl. Takoj po prvem dodatku je okno odprto – zavezal se je eni stvari in je za kratek čas dovzeten za to, kar sodi zraven.

Če to okno zapraviš z generično mrežo “morda ti bo tudi všeč”, dobiš temu ustrezno stopnjo pripenjanja: skoraj nič, ker blok očitno ne govori o njegovi izbiri. Priporočilo mora vidno povezano s tem, kar je pravkar izbral, sicer ga bere kot tapeto.

Tri vrste izdelkov, ki jih je vredno pokazati

Skoraj vsako dobro priporočilo po dodatku sodi v enega od treh košev. Preden posežeš po čemer koli drugem, izberi iz teh.

1. Tisto, kar potrebuje, da deluje. Baterije za igračo, črnilo za tiskalnik, kartica SD za fotoaparat, nosilec za polico. Ti imajo najvišjo stopnjo pripenjanja, ker jih bo kupec tako ali tako potreboval, in če jih ne pokažeš, mu skoraj delaš medvedjo uslugo. Če doma ugotovi, da fotoaparat nima priložene pomnilniške kartice, je to vračilo ali slaba ocena, ki že čaka.

2. Tisto, kar dokonča delo. Nekoliko mehkeje od stroge zahteve, a dokonča, kar je kupec začel. Kupil je tekaške čevlje? Tekaške nogavice in pas z ustekleničeno vodo. Kupil je litoželezno ponev? Olje za utrjevanje in strgalo. Tu spremeniš en izdelek v rešitev – celoten pristop sem posebej opisal v kako iz naročila z enim izdelkom narediti celovito rešitev.

3. Tisto, kar so ljudje, ki so to kupili, dejansko kupili zraven. Kadar dodatek ni očiten, resnični podatki o nakupih premagajo tvojo intuicijo. Če zgodovina naročil kaže, da kupci izdelka A pogosto kupijo tudi izdelek C – čeprav ju sam nikoli ne bi združil – zaupaj podatkom. Iskanje teh parov je posebna veščina; kako najti kombinacije izdelkov, ki jih kupci naravno kupujejo skupaj te vodi skoznjo.

Začni s prvim košem. Šele ko izdelek nima očitnega spremljevalca, se pomakni na drugega in tretjega – in če nima prav nobenega, kaj priporočiti, ko izdelek nima očitnega dodatka obravnava natanko ta primer.

Zakaj običajna mreža “morda ti bo tudi všeč” tiho odpove

Večina trgovin spusti noter gradnik s priporočili, ga nastavi na “priljubljeni izdelki” ali “sorodno po kategoriji” in reče, da je opravljeno. Dve težavi.

Prvič, “sorodno po kategoriji” ni isto kot “gre skupaj”. Kupec, ki kupuje rdečo obleko, noče videti štirih drugih rdečih oblek – to so nadomestki, alternative stvari, ki jo je že izbral. Pokaži nadomestke po dodatku v košarico in povabiš k ponovnemu premisleku: mogoče sem izbral napačno. Kupca lahko dejansko odgovoriš od naročila. Dopolnila povečajo košarico; nadomestki krčijo prepričanost.

Drugič, “priljubljeni izdelki” prezrejo edini podatek, ki si ga pravkar prejel – kaj je dodal. Priporočilo, ki bi izgledalo enako ne glede na to, kaj je v košarici, ni priporočilo. Je pasica.

Razliko v stopnji pripenjanja sem videl na lastne oči. Generičen blok uspešnic po dodatku v košarico komajda registrira. Zamenjaj ga z dvema resnično dopolnjujočima izdelkoma, vezanima na vsebino košarice, in stopnja pripenjanja se povzpne v razpon, ki je vreden truda. Izdelki niso postali boljši. Relevantnost je.

Kje dejansko tiči izgubljen prihodek

Naredi grobo matematiko za svojo trgovino. Recimo, da 500 kupcev na mesec doda primarni izdelek za 40 €, ki ima naraven spremljevalni izdelek za 15 €. (Ilustrativno – uporabi svoje številke.)

Generična mreža, ~2 % pripenjanja: 10 spremljevalcev, 150 €/mesec.
Relevantno priporočilo, ki upošteva košarico, ~12 % pripenjanja: 60 spremljevalcev, 900 €/mesec – približno 9.000 € na leto.

Enak promet, enaka primarna prodaja, brez popusta. Edina spremenljivka je, ali je bilo priporočilo o njihovi izbiri ali o tvoji zalogi. Ta šestkratna razlika je denar, ki ga večina trgovin pusti ležati v privzetih nastavitvah gradnika, ki ga ni nihče uglasil.

Kako to avtomatizirati

To teče kot vedenjski sprožilec na spletni strani in mora delovati takoj.

Sprožilec: prvi izdelek, dodan v košarico v seji.
Segment: po dodanem izdelku (ali njegovi kategoriji), tako da predlog vedno odraža košarico. Prvi kupec in ponovni kupec lahko vidita tudi drugačno uokvirjanje – kako to razdeliti, je v kako segmentirati ponudbe navzkrižne prodaje po kupčevem prvem nakupu.
Umestitev: predal košarice ali majhen modul na strani – ne celozaslonski vmesni zaslon, ki blokira potek. Miren, mogoče ga je preskočiti, očitno neobvezen.
Vsebina: en ali dva dopolnjujoča izdelka z jasno oznako, kot je “gre s tem” ali “verjetno boš potreboval”. Vključi ceno. Naj bo dodajanje en sam dotik.
Varovalo: nikoli ne kaži nadomestkov, nikoli ne kaži več kot dveh ali treh, nikoli ne blokiraj poti do zaključka nakupa.
Cilj: stopnja pripenjanja priporočenega izdelka in višja povprečna vrednost naročila, brez padca dokončanih zaključkov nakupa.

Za časovno občutljive potrošne izdelke lahko to pozneje razširiš v opomnik po nakupu, a trenutek med sejo je tisti, kjer je prva, najcenejša zmaga.

Metrike, ki ti povedo, da deluje

Stopnja pripenjanja priporočenega izdelka – neposredni semafor za to konkretno mesto.
Povprečna vrednost naročila za seje, ki so videle priporočilo, v primerjavi s tistimi, ki ga niso – izolira dvig.
Dokončanje zaključka nakupa – varnostno preverjanje. Če pade, tvoj modul stoji na poti; naredi ga lažjega.
Stopnja vračil primarnega izdelka – neopazna. Če priporočanje pravega dodatka zniža vračila (manj pritožb “ni prišlo s tem, kar sem potreboval”), je to resničen dobiček, ki ga številka povprečne vrednosti naročila ne bo pokazala.

Če stopnja pripenjanja ostaja trmasto nizka celo potem, ko si priporočila vezal na vsebino košarice, se je težava premaknila od tega, kaj kažeš, k temu, kako in kdaj – vredno je prebrati zakaj večina ponudb navzkrižne prodaje ne deluje, čeprav so izdelki povezani, preden nadaljuješ z uglaševanjem.

Kje se vključi Omnisend

Logika priporočil – “ko se doda to, predlagaj tole” – je tisti zahtevni del za ročno napeljavo. V lastnih trgovinah uporabljam Omnisend deloma zato, ker njegove funkcije za priporočanje izdelkov in segmentiranje potegnejo dodani izdelek in ga ujamejo z dopolnjujočimi izdelki brez razvoja po meri, in ker isti podatki o kupcih poganjajo tako spodbudo na spletni strani kot morebitni nadaljnji stik po nakupu, tako da si oba ne nasprotujeta. Izbral sem ga pred Klaviyom, potem ko sem preizkusil oba, predvsem zaradi enostavnejše nastavitve ter združene e-pošte in SMS-a. Omnisend je partner Shopimationa v pridruženem programu; brezplačna različica pokrije dovolj, da to zgradiš in izmeriš, preden karkoli porabiš.

Orodje ujema izdelke. Ne more ti povedati, katera povezava resnično pomaga kupcu – ta presoja in poznavanje razlike med nadomestkom in dopolnilom ostajata tvoja.

Tvoj naslednji korak

Vzemi svoje tri izdelke z največjim obsegom prodaje in za vsakega zapiši en izdelek, ki bi ga kupcu za pultom podal dober prodajalec. To je tvoj prvi nabor priporočil. Ta teden napelji te tri kot predloge, ki upoštevajo košarico, nato po desetih dneh preveri stopnjo pripenjanja in dokončanja. Ko boš pripravljen odločiti, ali te spremljevalce prikazati posamično ali kot komplet s popustom, je komplet ali posamezna priporočila: kaj bolj dvigne vrednost naročila naslednja odločitev.

Average Order Value Growth

What to Recommend After a Customer Adds the First Product to Cart

Posted on July 18, 2026 by

The moment someone adds their first product to cart, recommend the item that makes that product work — the accessory it needs, the consumable it runs on, or the piece that completes the job they came to do. Not a bestseller. Not a bigger version of the same thing. Not six random products. One or two things that a knowledgeable shop assistant would hand them at the counter, because leaving without them would mean coming back annoyed. That first add-to-cart is a signal of intent you rarely get this cleanly, and what you show next either feels like help or like a sales pitch. This page is about exactly what to put in that slot — which products, in which order, and what to avoid.

Why this moment is different from every other recommendation

A first add-to-cart tells you three things at once: what the customer wants, that they’re ready to buy, and that they’re still in a shopping mindset rather than a paying one. That combination is rare. On the homepage you’re guessing. At checkout they’ve mentally closed the decision. Right after the first add, the window is open — they’ve committed to one thing and are briefly receptive to what goes with it.

Waste that window on a generic “you might also like” grid and you get the attach rate to match: near zero, because the block clearly isn’t about their choice. The recommendation has to visibly connect to the thing they just picked, or it reads as wallpaper.

The three types of item worth showing

Almost every good post-add recommendation falls into one of three buckets. Pick from these before anything else.

1. The thing it needs to work. The batteries for the toy, the ink for the printer, the SD card for the camera, the mounting bracket for the shelf. These have the highest attach rate because the customer will need them anyway, and not showing them is arguably a disservice. If they find out at home that the camera doesn’t include a memory card, that’s a return or a bad review waiting to happen.

2. The thing that completes the job. Slightly softer than a strict requirement, but it finishes what they started. Bought running shoes? Running socks and a water belt. Bought a cast-iron pan? The seasoning oil and a scraper. This is where you turn a single item into a solution — a whole approach I’ve written up separately in how to turn a single item order into a complete solution.

3. The thing people who bought this actually bought too. When the accessory isn’t obvious, real purchase data beats your intuition. If your order history shows that buyers of product A frequently also buy product C — even if you’d never have paired them — trust the data. Finding those pairs is its own skill; how to find the product combinations customers naturally buy together walks through it.

Start with bucket one. Only when a product has no obvious companion do you move to two and three — and if it has none at all, what to recommend when a product has no obvious accessory covers that case directly.

Why the usual “you might also like” grid quietly fails

Most stores drop in a recommendation widget, set it to “popular products” or “related by category,” and call it done. Two problems.

First, “related by category” is not the same as “goes together.” A customer buying a red dress does not want to see four other red dresses — those are substitutes, alternatives to the thing they already chose. Show substitutes after add-to-cart and you invite second-guessing: maybe I picked the wrong one. You can actually talk them out of the order. Complements grow the basket; substitutes shrink confidence.

Second, “popular products” ignores the one piece of information you just received — what they added. A recommendation that would look identical no matter what’s in the cart isn’t a recommendation. It’s a banner.

I’ve seen the attach rate difference first-hand. A generic bestseller block after add-to-cart barely registers. Swap it for two genuinely complementary items tied to the cart contents and the attach rate climbs into a range worth the effort. The products didn’t get better. The relevance did.

Where the lost revenue actually sits

Run the rough math on your own store. Say 500 customers a month add a €40 primary product that has a natural €15 companion. (Illustrative — use your numbers.)

Generic grid, ~2% attach: 10 companions, €150/month.
Relevant, cart-aware recommendation, ~12% attach: 60 companions, €900/month — about €9,000 a year.

Same traffic, same primary sales, no discount. The only variable is whether the recommendation was about their choice or about your inventory. That six-fold gap is the money most stores leave sitting in the default settings of a widget nobody tuned.

How to automate it

This runs as an on-site behavioral trigger, and it should feel instant.

Trigger: first product added to cart in the session.
Segment: by the added product (or its category), so the suggestion always reflects the cart. A first-time buyer and a repeat buyer might also see different framing — how to split that is in how to segment cross-sell offers by the customer’s first purchase.
Placement: the cart drawer or a small on-page module — not a full-screen interstitial that blocks the flow. Calm, skippable, obviously optional.
Content: one or two complementary items with a plain label like “goes with this” or “you’ll probably need.” Include price. Make adding them a single tap.
Guardrail: never show substitutes, never show more than two or three, never block progress to checkout.
Goal: attach rate on the recommended item and a higher AOV, with no drop in checkout completion.

For time-sensitive consumables you can extend this into a post-purchase reminder later, but the in-session moment is where the first, cheapest win is.

The metrics that tell you it’s landing

Attach rate on the recommended item — the direct scoreboard for this specific slot.
AOV for sessions that saw the recommendation vs. those that didn’t — isolates the lift.
Checkout completion — the safety check. If it dips, your module is getting in the way; make it lighter.
Return rate on the primary product — a subtle one. If recommending the right accessory drops returns (fewer “it didn’t come with what I needed” complaints), that’s real profit the AOV number won’t show.

If the attach rate stays stubbornly low even after you’ve tied recommendations to cart contents, the problem has moved from what you show to how and when — worth reading why most cross-sell offers fail even when the products are related before you keep tuning.

Where Omnisend fits

The recommendation logic — “when this is added, suggest these” — is the fiddly part to wire up by hand. I use Omnisend in my own stores partly because its product recommendation and segmentation features pull the added product and match it to complementary items without custom development, and because the same customer data drives both the on-site nudge and any post-purchase follow-up, so the two don’t contradict each other. I chose it over Klaviyo after testing both, mainly for the simpler setup and combined email and SMS. Omnisend is an affiliate partner of Shopimation; the free tier covers enough to build and measure this before spending anything.

The tool matches products. It can’t tell you which pairing genuinely helps the customer — that judgment, and knowing a substitute from a complement, stays yours.

Your next step

Take your three highest-volume products and, for each, write down the single item a good shop assistant would hand the buyer at the counter. That’s your first recommendation set. Wire those three up as cart-aware suggestions this week, then check attach rate and completion after ten days. When you’re ready to decide whether to show those companions individually or as a discounted bundle, bundles or individual recommendations: which raises order value more is the next call.