A driver column contains a metric that is expected to have a (more or less) linear relationship with the cost/revenue column, everything else being equal.
These are columns such as Visits, Bookings, Distribution.
Drivers are metrics (distribution, visits, checkouts) that have some direct impact on sales and that therefore help explain the change in sales.
For instance if a company sells in one shop only and sells 100$ in year 1, one might assume that if in year 2 it sells 200$ and has added a second shop, the increase is due to the larger distribution - assuming every thing else stays equal.
Similarly if the number of visits to the site doubles and everything else (conversion rate, average price, volumes per purchase act) stays the same, you would expect revenues also to double. If the sales of a website double and the visits to the website double - assuming every thing else stays equal - that increase in due to more visitors. If the visitors stay the same but the number of checkouts double along with the sales, that increase is due to a better conversion ratio.
These metrics have to be in value (not in %) - for instance, number of booking not conversion rate.
Note that the dataset above contains no ratios, just the raw values of Visits, Bookings, Nights and Amount. Bookings on Visits is the conversion ratio. Nights on Bookings is the average number of nights of stay.
'Mparanza returns the variance - the impact on sales in other words - tied to the change in the number of visits, in the proportion of visitors that buy, in the number of outlets that stock the product.
'Mparanza calculates driver variance, making it easy to understand what is driving change.
If the dataset has one or more driver columns the tool will return one or more "driver variances", that maps the impact/s of the driver/s.
The tool manages the following driver columns:
- Category weighted distribution. Measures the weighted number of outlets where a given product is available for sales. If distribution points double, sales double and everything else stays the same, the increase will be allocated to distribution variance rather than volume variance (which would measure the "per outlet" sales change).
- Checkouts. Measures the number of purchase decisions (tickets, bookings, checked out shopping carts,...). If checkouts double, sales double and everything else stays the same, the increase will be allocated to checkouts variance rather than volume variance. Everything equal an increase in check out variance indicates an improvement in conversion.
- Visits. Measures the number of customer interactions (shop/website visits, traffic, inquiries, leads,...). If visits double, sales double and everything else stays the same, the increase will be allocated to visits variance rather than volume variance. Everything equal an increase in visits variance indicates an improvement in traffic.
If your dataset has other driver columns you would like us to include in the tool, please contact us.