Why do you need an analytics culture?

The root of the problems that can arise when moving to flat teams lies in communication. In a company where one person makes the decisions, the question of analytical culture simply does not arise. In such companies, the manager has their own vision and assigns tasks to subordinates on their own. Those, in turn, pass tasks down to their employees - and so on.

This is how the companies we are used to, with a hierarchy of positions, operate.

The weak point of this kind of setup is shifting responsibility.

The subordinate, being closest to the customer, production, and/or reality, realizes that solving the task in this way is irrational. 'But if the boss said it must be done, then it must be done.'

And if the end result is poor, then the manager is to blame, since they set the task in the first place. There is another unpleasant aspect to the hierarchical management model, this time a personal one. As someone who strives to grow, I want to be around strong people and relate to them as equals.

But when the responsibility lies with the person assigning the task, they have to issue instructions and tell others what to do in order to achieve the desired result.

The person doing the work loses the opportunity to express their opinion or their idea. Of course, in such a case there can be no question of equal communication.

Flat teams emerge where the person doing the work does not follow the task strictly according to the technical specification.

They see the problem and figure out for themselves what needs to be done and how.

Everyone only agrees on success metrics.

I often hear managers express one concern about the move to flat teams - 'there will be too many points of view, and everyone will pull in their own direction.'

Or another concern (which is essentially the same thing) - that responsibility will be distributed very vaguely among employees, and they will simply start passing it to one another.

Such concerns are not unfounded, but the move to flat teams still looks very attractive - after all, a company that follows this concept can live and grow without a leader. And this is not only about the elegance of the ethics.

Getting out of day-to-day operations, no longer being tied to workflows, and finally finding time for ideas and strategic planning is probably every leader's wish.

The root of the problems that can arise when moving to flat teams lies in communication.

If you can no longer give orders and hide behind your senior position, how do you justify your own point of view?

If age, tenure, and experience are no longer the key factors in deciding whose opinion to listen to, how can a team tell whose view is truly interesting, useful, and valuable? And if you imagine that it is possible to step out of day-to-day operations and not control which tasks the team is handling, then how do you keep the right development direction and avoid chaos?

To move communication out of the realm of opinions and into the level of reasoned dialogue, and to turn personality clashes into conflicts of interest, you need to learn to rely on facts.

A fact is something everyone agrees on, something that does not depend on point of view. For example, the number of sales or unique sessions on a website.

The average number of items per customer will be approximately equal to

Let us imagine that the stage of being infused with analytical culture has been completed, and each department or counterparty has found its own metrics.

Everyone works, shows initiative, and achieves the goals set. But there is no result. Or there is one, but not the right one.

If departments in a company may not know what is happening with their colleagues in other departments, the customer, by contrast, moving from the moment a need arises to recommending the company to neighbors, comes into contact - directly or indirectly - with the work of every department, counterparty, and/or contractor. The points where one party's responsibility ends and another's begins are sometimes not tracked at all.

Accounting tracks its metrics in 1C, sales in CRM, and marketers in Google Analytics. Each of these systems may be fine on its own. But to see the full picture, you need an integrator that enables end-to-end analytics.

BI systems can be used to solve this issue. BI systems are systems that aggregate information from different sources.

The dashboard shows how many orders there were in total, how many were moved to the warehouse, how many were delivered to the end customer, and how many were canceled (Fig. 1).

Order Collection and Delivery Services System

If more detailed data is needed, we select canceled_lms.

A list of the errors that occurred will be displayed below the dashboard (Fig. 2).

A detailed breakdown of the errors that occurred. In this case, an understanding of how the order-picking-delivery chain works.

, which errors occur and when helps determine whether anything else needs to be refined in the integration or whether it is already ready for launch.

At first, half of the orders were canceled at the delivery stage; now it is around 7%.

The problem was that the addresses the online store passed on to delivery could not be used to create orders. In some cases the region was missing, and there are about a dozen cities with the same name across the country; in others, the street name was entered incorrectly, and so on.

It was important to understand at which stage orders were being lost and then agree with the online store on optimizing the processes for capturing, storing, and processing customer addresses.

But most importantly, all stakeholders now see the same data and can agree on metrics, ways to solve problems, and the project's future development. With BI, we were able to set up not only tracking of errors as they arise, but also the collection of detailed information about why they occur, which ultimately made it possible to make a number of important decisions about the project's future development.