Understand how distinct customer cohorts buy, repeat and churn
A single "average customer" figure is a poor guide when your business serves more than one type of buyer. If trade accounts ordering weekly, trade-counter collection buyers and one-off retail shoppers are all rolled into one store-wide average, your reports will hide who actually drives your margin and who never comes back. - Store-wide conversion or Average Order Value looks steady, while repeat trade orders are quietly slipping. - Marketing judges ad campaigns on first-order value alone, cutting channels that bring in repeat trade accounts. - Automated "VIP" or win-back discount emails get sent to contracted trade accounts that were going to reorder anyway. - Guest checkouts and multiple buyers at the same company make repeat customers look like one-off visitors in analytics. Agentix provides customer, cohort and repeat-purchase analysis for UK eCommerce, B2B and trade businesses. Combining our [customer data and systems](/services/customer-data-systems) and [website conversion](/services/website-conversion) expertise, we separate your customer base into meaningful cohorts using your actual order records so you can make clearer decisions on acquisition spend, retention and site journeys.
How a blended store average hides your real customer economics
In our experience, looking at a single blended Average Order Value or repeat purchase rate across a whole store leads teams to invest in the wrong places. Here is an illustrative example with invented round numbers (not a client result) showing a business turning over **£4,432,500 across 10,000 customers and 18,500 orders**, comparing the blended dashboard view against the separated cohorts underneath: | Customer cohort | Customers (% of total) | Annual orders | Average Order Value | Annual repeat rate | Annual revenue (% of total) | What the blended average hides | | -- | -- | -- | -- | -- | -- | -- | | **Blended store total** | **10,000 (100%)** | **18,500** | **£240** | **32%** | **£4,432,500 (100%)** | **Describes nobody accurately; mixes £85 retail shoppers with £480 trade accounts.** | | 1. One-off retail consumers | 7,200 (72%) | 8,100 | £85 | 11% | £688,500 (15.5%) | High volume of customers, but nearly 9 in 10 never buy a second time. | | 2. Managed trade accounts | 800 (8%) | 6,800 | £480 | 84% (8.5 orders/yr) | £3,264,000 (73.6%) | Just 8% of accounts generate nearly three-quarters of total revenue. | | 3. Collection / trade counter | 1,200 (12%) | 2,400 | £145 | 42% (
Where customer analytics breaks down in practice
When customer segments are built from browser cookies or unlinked records rather than clean order history, the same commercial mistakes show up repeatedly: | What you see in your reports | What is happening underneath | What it costs your decisions and margin | | -- | -- | -- | | Overall retention looks stable at 30%+, yet cash flow from repeat orders is tightening | Long-standing trade accounts are propping up the store average while newly acquired cohorts have stopped repeating | You keep funding acquisition campaigns whose new customers churn after a single order | | Profitable B2B campaigns get paused because first-order ROAS looks too low | Campaigns are judged on the initial trial order rather than 6-month or 12-month account value | You cut the exact channels that bring in your highest-value trade and commercial accounts | | Trade accounts receive 15% "we miss you" discount vouchers | Win-back rules use a single 60-day inactivity timer without checking seasonal buying cycles or contracted trade pricing | You give away 15% margin on orders that contracted accounts were already due to place at full price | | Repeat customer rate in GA4 looks far lower than in your finance sys
Defining whether a "customer" is a person, an account or a site
Before we draw a single chart or cohort table, we check how your records identify a customer: - **Individual Contact vs Company Account vs Delivery Location:** If three site managers order for the same contractor, counting them as three separate customers will distort both your average account value and your retention rate. - **Order ledger truth over browser cookies:** Because browser cookies expire and buyers switch devices, we run cohort and repeat-purchase analysis against your actual order records (Shopify, Magento, ERP or CRM) rather than relying on web analytics sessions alone. - **Explicit visibility for unknown / guest orders:** Where guest purchases cannot be linked reliably to a known account, we keep them visible as a separate group rather than quietly folding them into retail or trade totals.
Frequently asked questions
### Why shouldn't we rely on GA4 for customer retention and lifetime value? GA4 tracks browser cookies unless a visitor logs into an authenticated session on every visit. When B2B or trade buyers switch between a desktop PC in the office and a mobile phone on site, clear their cookies, or place an order by phone, GA4 counts them as brand-new users. We use your transactional order database and CRM records so repeat purchasing is measured against real customers and company accounts. ### How do you handle businesses that sell to both trade accounts and the general public? We separate them before calculating any averages. Trade accounts, trade-counter collection buyers and retail consumers have completely different order values, margins and buying frequencies. Analysing each cohort on its own stops high-volume retail orders from masking what is happening in your trade accounts. ### Can you analyse repeat purchasing if many of our customers check out as guests? Yes. We check whether guest orders can be linked deterministically using verified billing emails, phone numbers or delivery details. Where a link is clear, we group those orders accurately; where identity remains uncertain, we
Talk to us about customer and cohort analysis
Whether you want to separate trade and retail performance, understand your true repeat-purchase cycles, or find out which acquisition channels bring in customers who stick around, we can help.