Skip to content
  • Published
  • Furniture and chairs
  • Poland

Case study - Furniture e-commerce

Meta missed 41.6% of the orders its own ads started. AxiTrace found every one.

CentrumKrzesel.pl sells chairs, tables and sofas online and spends over $300,000 a month on Google, Meta and Microsoft ads. Their reports credited the last click. AxiTrace showed the campaign that actually started every journey, and it changed which campaigns look worth scaling.

  • Free up to 10,000 visits a month
  • No credit card
  • Live the same day
Dining chairs around an oak table in a dimly lit furniture showroom

The results in numbers

  • 41.6%

    of Meta-started orders came after day 7

    outside Meta's default 7-day click window; 14.3% after day 30

  • +35%

    revenue credited to non-brand Google campaigns

    more revenue under first click than last click gave them

  • 26%

    of all ad-started orders took 7+ days

    from the first ad click to the order, up to 73 days

  • < 24 h

    from kickoff to full tracking

    first order tracked 5 hours after setup started

Measured on 20,000+ real orders. Test orders excluded. Source: the store's AxiTrace workspace.

Customer snapshot

CentrumKrzesel.pl

A furniture store spending over $300,000 a month on ads saw which campaign really started each order, even weeks later, and found non-brand Google campaigns started 35% more revenue than last click showed.

Industry

Furniture and chairs for home and office

Market

Poland

Store

Custom-built platform, no plugin

Ad channels

Google Ads (Search, Shopping, Performance Max), Meta, Microsoft Ads, OpenAI Ads, price comparison sites, display retargeting

Ad spend

Over $300,000 a month

Store scale

20,000+ orders and 20M+ in revenue measured

Average order

1,400+

AxiTrace modules

Attribution, unified ads dashboard, Fraud Detection, Axi AI assistant

Visit the store

The challenge

Nobody buys a dining set on the first click.

CentrumKrzesel.pl is one of the largest online furniture stores in Poland: office chairs, dining chairs, tables, sofas and armchairs, with an average order above 1,400. Half of the buyers decide the same day. The other half compare models, fabrics and prices for days, come back through brand search, a price comparison site or a newsletter, and only then order.

Every ad platform reports that journey through its own keyhole. Google Ads and Meta count conversions inside their own click windows and only for their own clicks. A last-click report gives the whole order to the final visit, which for a considered purchase is usually brand search, organic search or an email. The campaign that found the customer two weeks earlier gets nothing.

At over $300,000 a month across more than 100 campaigns, that is not a reporting detail. It decides which campaigns get scaled and which get cut.

A leather armchair, an oak dining chair and an office chair side by side
  • 63.5%

    of orders needed 2 or more visits

    one in four needed 5 or more

  • 4.0

    tracked visits per order on average

    across ads, search, email and direct

  • 100+

    campaigns with spend

    on Google Ads, Meta, Microsoft Ads and OpenAI Ads

Long purchase decisions

Half buy today. The rest take days, weeks, even months.

Time from the first paid ad click to the order, for every order that started with an ad. Orders that took one to four weeks were also the bigger baskets.

Days from the first ad click to the order

  1. Same day

    baseline order value

    51.8%

  2. 1-2 days

    orders 22% larger

    10.9%

  3. 3-6 days

    orders 24% larger

    11.2%

  4. 7-13 days

    orders 28% larger

    10.2%

  5. 14-29 days

    orders 28% larger

    9.2%

  6. 30+ days

    orders 21% larger

    6.8%

Share of ad-started orders. Under each bucket: average order value compared with same-day orders.

  • 26.2% of ad-started orders were placed 7 or more days after the first ad click, and they brought 29.5% of that revenue.
  • 18.4% took more than 12 days. The longest journey in the data took 73 days.
  • Buyers who took 7 to 29 days spent 28% more per order than same-day buyers.

The Meta blind spot

4 in 10 Meta-started orders arrived after Meta stopped counting.

Meta credits a purchase by default only if it happens within 7 days of the ad click. CentrumKrzesel.pl's buyers take longer. Of all orders that started with a click on a Meta ad, these came after the window closed:

41.6%

of Meta-started orders came after day 7

44.7% of all the revenue that Meta ads started

Days from the first Meta ad click to the order

  1. Same day

    34.4%

  2. 1-2 days

    11.4%

  3. 3-6 days

    12.6%

  4. 7-13 days

    14.0%

  5. 14-29 days

    13.3%

  6. 30+ days

    14.3%

Meta default 7-day click window ends

Share of Meta-started orders by days from the first Meta ad click. AxiTrace counts the whole journey, whatever its length.

  • 14.3%

    came after day 30

    15.5% of Meta-started revenue

  • 4.25 days

    median from Meta click to order

    Google-started orders: 0.67 days

  • 72 days

    longest Meta-started journey

    from the first click to the order

Who really started the sale

Last click paid the closer. First click found the campaigns that open the door.

The same orders, credited two ways: to the campaign of the first click and to the campaign of the last click. Revenue as an index, last click = 100.

Revenue credited under first click vs last click

  • Non-brand Google campaigns

    Shopping, Performance Max and generic search, 90 campaigns

    +35%

    First click

    135

    Last click

    100

  • Meta catalogue prospecting

    Advantage+ catalogue campaign for new customers

    +40%

    First click

    140

    Last click

    100

  • Brand search

    Google Ads campaign on the store's own name

    -23%

    First click

    77

    Last click

    100

  • Display retargeting

    Retargeting network, shown after a visit

    -59%

    First click

    41

    Last click

    100

Orders with a paid first click, 26 August to 9 November 2026. Index per row: revenue under last click = 100.

  • 29.7% of ad-started orders ended on a different source or campaign than the one that started them, carrying 31% of their revenue.
  • Of the 34 non-brand Google campaigns with at least 30 first-click orders, 30 earned more under first click, and 14 earned 50% more or better. One Shopping campaign started 3.1 times the revenue that last click gave it.
  • Brand search and retargeting do real work at the end of the journey, but under last click they collect credit for customers that other campaigns found.

One real order

24 days, 48 visits, one chair.

A real order from the data, without customer details. A last-click report gives this chair to brand search. AxiTrace shows the Meta catalogue ad that started it.

  1. Day 0

    Meta catalogue ad

    First click, prospecting campaign

    1 visit

  2. From day 4

    Meta ads again

    Prospecting and retargeting

    16 visits

  3. From day 11

    Google Performance Max, Google app, a price comparison site

    Comparing models and prices

    4 visits

  4. From day 16

    Direct, ChatGPT, Facebook, social, display retargeting

    Coming back to decide

    15 visits

  5. From day 17

    Google Ads

    Performance Max, brand search and one more campaign

    11 visits

  6. Day 24

    Google brand search

    Last click before the order

    1 visit

  7. Day 24

    Order

    One chair

Visits grouped by stage. Every visit was tracked in the store's AxiTrace workspace.

  • Meta

    started the journey

    credited under first click

  • Brand search

    got the last click

    credited under last click

  • 9

    different sources on the way

    ads, search, price comparison, AI, social and direct

The implementation

Custom store, no plugin. Live in under 24 hours.

CentrumKrzesel.pl runs on its own e-commerce platform, so there was no app to install. Their developers added the AxiTrace browser SDK and sent orders from the server, following the step-by-step documentation.

  1. Day 1, early afternoon

    Workspace created

    Tracking key issued and the documentation handed to the store's developers.

  2. Day 1, evening

    Live on the whole store

    Page views, product views and carts flowing for every visitor. The first order was tracked 5 hours after the workspace was created.

  3. Day 2

    Ad spend connected

    Google Ads and Meta accounts connected read-only, so every campaign shows cost next to the orders it started.

  4. Week 6

    Fraud Detection switched on

    Suspicious IP addresses excluded from Google Ads automatically, with the store's confirmation.

  5. Months 2-3

    More channels, one model

    Microsoft Ads and OpenAI Ads joined the same dashboard. After two months of data the marketing team made first click the store's default attribution model.

Our store runs on its own platform, so there was no plugin to click. It did not matter. The documentation was precise, every event was described with an example, and we had tracking, orders and ad spend live in under 24 hours.

E-commerce team
CentrumKrzesel.pl

Read the integration docs

Customer story

CentrumKrzesel.pl

Our customers think about a chair or a dining set for days, sometimes for weeks. Google and Meta only showed us the last click, so the campaigns that brought a customer in for the first time looked weak, and brand search looked like a hero. With budgets like ours you cannot scale on a hunch. AxiTrace shows us which campaign really started each order, and that is now how we decide where the next zloty goes.

Owner
CentrumKrzesel.pl
A velvet dining chair lit by a warm lamp

What the AI assistant found

A morning briefing that pays for itself.

Axi, the AxiTrace AI assistant, reads the store's orders, campaigns, ad spend and live Google Ads data, and starts every day with a short briefing. These are real answers it gave the CentrumKrzesel.pl team.

  • Performance Max

    Budget up 51%, revenue down 59%

    One Performance Max campaign spent 51% more week on week while its revenue fell 59%. Axi checked Google Ads live: cost per click had not changed, Google had bought about 36% more impressions, and almost all of the extra spend went to generic, uncategorised search queries.

    A ROAS drop of almost three quarters, explained in one answer

  • Search terms

    A ready negative keyword list

    Over 60 days six generic queries took 4.1% of that campaign's budget and brought almost no conversion value. Axi wrote the exact-match exclusion list, and corrected itself on "ergonomic" queries, which sell well.

    Budget with almost no sales, found and listed

  • Cannibalisation

    Two campaigns, the same chairs

    The weakest Google campaign, with a ROAS about a third of the account average, shared its two best-selling chairs with a broader Performance Max campaign that kept spending on them too. The products were never excluded when the new campaign started.

    Overlap found before cutting the campaign

  • Retargeting

    "Should we switch it off?"

    Asked about a display retargeting network, Axi showed all three views: last click gave it almost four times as many orders as first click, and it appeared in three times more journeys still. It almost never brings a new customer on its own, so Axi advised a test: pause or halve it for 2-3 weeks and watch total sales before cutting.

    A decision backed by three attribution views

Almost every day Axi tells us something we had not spotted: a Performance Max campaign buying generic queries, two campaigns bidding on the same chairs, a budget that grew while revenue fell. It is like having an analyst who went through every campaign, segment and search term overnight and left us a short note with the numbers.

Performance marketing
CentrumKrzesel.pl

Figures as Axi reported them from the store's data and Google Ads, November 2026. Axi reads data and recommends; it never changes the ad accounts.

How Axi and the AI Advisor work

Also on the same data

Paid clicks that were never going to buy, filtered out.

Fraud Detection runs on the same tracked visits. With the store's confirmation it excludes suspicious IP addresses in Google Ads automatically. Last 30 days:

  • 1,000+

    paid sessions flagged

    from 200+ suspicious addresses

  • 100+

    IP exclusions pushed to Google Ads

    automatically, after the store confirmed

  • 100,000+

    bot hits filtered

    before they could reach any report

About Fraud Detection

How we measured

Every number on this page comes from the store's own AxiTrace data. One period and one method for every figure.

  • Period: from 26 August 2026, the first tracked order, to 9 November 2026.
  • Excluded: test orders placed by the store's development environment. Revenue is order value as sent by the store.
  • First click and last click: the paid or organic visit that started and that closed each journey, as tracked by AxiTrace across sessions of the same visitor, including IP fallback for visits without a cookie.
  • Time to purchase: hours from the first paid ad click to the order. Clicks made before tracking began are not visible, so long journeys are, if anything, undercounted.
  • Ad spend: imported from Google Ads, Meta, Microsoft Ads and OpenAI Ads at campaign level.

Find the campaigns that really start your sales.

Connect your store and ad accounts and see first click, last click and data-driven attribution side by side on your own orders. Free plan, no credit card.

Start Free

An agency, or more than 3M visits a month? Talk to us

  • Data stored in the EU
  • GDPR, with a DPA
  • On the Shopify App Store
  • Plugin on WordPress.org
  • In ChatGPT and Claude

AxiTrace demoSample store

Store revenue$42,300last 30 days
Ad spend$9,9755 campaigns on 3 ad platforms
Store ROAS4.24revenue over ad spend
Orders1,223from every source

ROAS next to profit after costs

Store ROAS4.24 Campaign profit+$3,015

Store ROAS counts all revenue. Campaign profit subtracts the product costs and the ad spend of each campaign.

Axi, morning briefing

Two campaigns lose money after costs: Google Shopping and TikTok Video. Meta Prospecting holds a steady profit and is the one to scale in steps.

Sample data. In your own account you see your own store.

Start Free