A sales director asks a simple question: “Which customers dropped the most this month?”
It sounds like a five-minute job. It rarely is. Someone has to pick the dates. They run one report for this month and another for the same month last year. They export both to Excel, line the numbers up by hand and check them twice. The answer lands in the inbox two or three days later. By then the month has moved on, and so have the customers.
If that sounds familiar, you are not alone. And the problem is not what most people think.
The paradox: lots of data, few answers
Most mid-to-large companies are not short of data. Years of sales history sit in the ERP(Enterprise Resource Planning). Customer records live in the CRM. Everything else ends up in a data warehouse someone built a long time ago. If you run a loyalty program, that is another system full of points, orders and member history.
All of it is there. The trouble is how hard it is to ask for anything. Reports are built for the questions someone thought of years ago. A new question means a new report, or a manual workaround. So companies end up rich in data and poor in answers.
You can see the cost in small ways. Managers stop asking questions because they know how long the wait is. Analysts spend their week rebuilding the same spreadsheet. A customer who quietly buys less gets noticed late, sometimes after they are gone.
What it looks like on a normal month
Recently, a client sent us several real questions their managers ask every month. Here are a few:
“Which distributors are selling less of our newest product line than expected?”
“Which products had zero sales last month, and in which regions?”
“What is this month’s sales total, and how does it compare with year-to-date?”
These are not strange questions. They are exactly what a good sales manager should be asking. But each answer today means picking dates, running two or more reports, exporting and stitching the results together by hand.
Who we are
We take care of large, long-running business platforms. Our biggest example is a 12-year-old enterprise platform that serves a global professional beauty brand’s Canadian business. Its users are the brand’s sales teams, distributors, educators and thousands of salons.
The platform includes sales reports and drill-downs, a customer database, a field tool for educators and a loyalty program with about 3,400 active members. It holds millions of rows of sales data, loaded every month from distributors, plus loyalty points, orders and education records.
We maintain it, improve it step by step, and connect it to the brand’s SAP system for things like loyalty orders, order status and stock.
Our approach: keep the old system, add a new layer
When a platform has worked for many years, the usual advice is “replace it”. We think that is often the wrong call. Old systems hold the history, the business rules and the trust of the people who use them. Throwing that away is expensive and risky.
So we do the opposite. We keep the system running. We modernize it one piece at a time. We connect it to the systems around it. And now we are adding something new on top: an AI assistant layer that lets people talk to their data in plain language.
Instead of choosing dates, filters and groupings, a manager types a question and gets a table, a chart and a short written summary. The assistant is designed to use the permissions and rules the platform already has, so it works with your system, not around it.
Proof that the data is worth digging for
Here is one number from that same platform. Loyalty members spend roughly 9x more per year than non-members. They also buy almost twice as many brands.
That is the kind of insight that can shape a whole sales strategy: where to invest, who to call, which customers to bring into the program. And it was sitting in the data all along. The only thing in the way was the effort needed to find it.
We would bet your data holds insights like this too. They are just waiting behind a few days of report-building.
What is next
We started with one simple question: what if you could just ask your data? Next week we will show what that looks like, with examples of asking, understanding and predicting.
Until then, a question for you: how long does it take your team to answer a simple sales question? Follow us to catch the rest of the series.