In this article
- Why a customer’s warning signs stay invisible when each team only sees its own tool.
- One question that queries three sources at once: the sales database, the CRM and the ticketing tool.
- The clarification the agent asks for before talking about a “drop”.
- Reminders created in the CRM, then follow-up through a dashboard and an alert.
The situation
At Nordica, a fictional B2B distribution company, three teams each see part of the same customer. Sales reps live in the CRM, where they follow their accounts, meetings and opportunities. Customer service lives in the ticketing tool, where late deliveries, billing errors and access problems land. Leadership looks at revenue, which comes out of the sales database.
None of these views is wrong. But a customer who orders less and has been waiting two weeks for an answer to a ticket sends a signal nobody sees in full. The sales rep notices the drop at the quarterly check-in; customer service handles the ticket without knowing the account is fragile. By the time someone puts the two together, the customer may already have decided.
The questions you would like to ask
Retention questions are simple to phrase, which is what makes them frustrating:
- Which customers cut their orders by more than 20% in a year?
- Which key accounts reduced their orders this quarter and have an open support ticket?
- How long have those tickets been open, and who owns the account?
- Which carrier is late most often?
The first one can be read from a single database. The second needs three. Without a tool to bring them together, it turns into a project: three exports, a mapping table between each system’s customer identifiers, and someone who spends the morning on it.
One question, three sources
Nordica’s head of customer service asks it as is. The databases, the CRM and the ticketing tool are all connected to the same workspace: the two business tools through their APIs, the sales database read-only.
Which key accounts reduced their orders this quarter and have an open support ticket?
- Sales databaseorders
- CRMaccounts, sales reps
- Ticketingopen tickets
Three key accounts combine falling orders with an open ticket:
| Account | Orders | Open ticket | Sales rep |
|---|---|---|---|
| Maison Castel | −18% | Late delivery | Claire Martin |
| Norvia | −12% | Billing error | Thomas Robert |
| Tessier et Fils | −9% | Portal access | Claire Martin |
The agent breaks the question down. The change in orders is computed in the sales database; the list of key accounts and their sales reps comes from the CRM; open tickets come from the ticketing tool. It queries each source for what it knows, then matches the results account by account. Nobody had to name the sources or build a mapping table.
A drop, compared with what?
A seemingly simple question often hides a choice. “Reduced its orders this quarter” can mean compared with the previous quarter or with the same quarter last year and, in a seasonal business, the two readings do not single out the same accounts. The agent does not decide for the team: it asks.
Period found: orders.order_date
Azul
To measure a drop in an account’s orders, I can compare the current quarter with the previous one, or with the same quarter last year. Which one best reflects a slowdown for you?
- The previous quarter
- The same quarter last year
- Both, side by side
Kept for the whole workspace
Nordica chooses the year-on-year comparison, which evens out seasonality. The choice is kept in the workspace’s memory: the next question about falling orders, whoever asks it, will use the same baseline.
Reading the list
Three accounts stand out, and the ticket column changes how the figures read. At Maison Castel, an 18% drop comes with a late delivery: the next question, asked in the same conversation, is about carriers. At Norvia, the ticket reports a billing error, and finance needs to be in the loop before any sales call. At Tessier et Fils, a portal access problem may well explain some of the orders that no longer come through.
The figure says an account is slipping. The ticket often says why.
Each account calls for a different action, but they all start the same way: the sales rep needs to know today, not at the next quarterly check-in.
Taking action
The head of customer service asks for the next step in the same conversation. The agent creates the tasks in the CRM, each assigned to the account’s sales rep and carrying the context of the ticket.
Create a follow-up task for each of them, assigned to their sales rep.
- Maison Castel: orders −18%Open ticket: late deliveryClaire Martin
- Norvia: orders −12%Open ticket: billing errorThomas Robert
- Tessier et Fils: orders −9%Open ticket: portal accessClaire Martin
Recorded in the audit log
That action is possible because Nordica connected its CRM and authorised task creation. Creating a ticket, updating a customer record, scheduling a follow-up: these only happen in the tools you connect, with only the actions you authorise, and each one is recorded in the audit log. The databases themselves stay read-only.
No more finding out after the fact
Today’s list only protects today’s accounts. For what comes next, the change in key-account orders is pinned to an “Accounts to watch” dashboard, refreshed every Monday at 7 am, before the week’s sales meeting.
An alert completes the follow-up. When a key account’s orders drop by more than 15% over the quarter, Azul e-mails the account’s sales rep and customer service, on the refresh that sees the threshold crossed. Nobody needs to open the dashboard to be in the know.
Threshold crossed: Maison Castel orders
- Measured
- −18%
- Threshold
- −15%
The refresh of the “Accounts to watch” dashboard found that Maison Castel’s orders fell by more than 15% over the quarter.
Open the dashboardFor the executive committee, the agent can also write a report on the quarter’s accounts at risk: its figures are frozen at the date it was written, so you can compare from one quarter to the next what was seen, and what was done.
What changes for the team
- A shared view of the customer. Orders, account and tickets are read together, without exports or mapping tables.
- The right person, straight away. The sales rep learns about the drop with its context, and customer service knows which tickets affect a fragile account.
- From finding to action in the same conversation. The follow-up lands in the CRM, assigned and traced.
- Vigilance that does not depend on anyone. The dashboard refreshes itself, and the alert speaks up as soon as an account crosses the threshold.
Where to start
Pick a segment that matters, key accounts for instance, and the question you would like asked every Monday. Read-only access to the sales database is enough to start; the CRM and the ticketing tool can be connected next, with only the actions you choose to authorise.