How a good business intelligence data helped me in the crisis.

Original post at: https://lab.mobi/articles/bi-for-crises

Current situation with the virus breakout is affecting business differently. But it is clear that there will be troubling times in economics. My company has been in business more than 19 years and I lived over the previous economic regression in 2008. I decided to give sone hands on ideas how to be prepared and have data available for hard decisions.

When I started working with my team, most of the attention went into finding people or finding clients. It does makes sense, as ultimately its the growth you need in the early days. You need to grow out from that freelancer mode of having few first people on the board. The same goes for the clients. Event though we have a globally successful client and cooperation, we always knew that we are way too dependant on them. But once you grow and are not able to micromanage every aspect of your business things start to get complicated. This starts usually earlier than you can predict.

In our case, the product design and development business, we are working with relatively inflexible resource – the people. You cannot buy them in for one week and let them go next one. But your clients expect you do help them fast. And when the client leaves or the virus makes your retail clients close the shop – you need some hard decisions to make.

In normal situation you can sell as much as you can and hope that the growth will save your ass. Or you could be prepared and start collecting some KPIs that will tell you which parts of the business needs improvement, which project is profitable and which one is not. If you are in the similar business as we are, following examples are probably not new to you:

• you design a shiny product but finish it for free because there is no budget to build it,
• you loose one big customer account that puts the whole business to risk,
• you might have understand that business is not profitable at the end of the year when doing annual accounting.

But after years of failing and trying, I believe we have become better when it comes to numbers. Now looking back, I believe it is possible to do it faster and I hope this article would save some time for you.

Step one – challenge your accountant

When it comes to accounting, there is something that has always bothered me. It is the fact that even though the balance sheet of the company and your income statement should describe the company accurately, it is rearly used as monthly tool to business management in small companies. In my opinion the system itself has created so many strict rules that in some point is started to drift away to wrong direction. So the accountant, instead of being the source for information to business manager, started to be a tool for tax department only. You should not pay for it, government should, right?

Luckily we have accounting team that has been with us for many years, they been with us for multiple spin-offs and start-ups we’ve created. So that allowed me to go to them and asked to fix three things:

• get the records in the income statement sheet reflect the structure of your business model
• get the invoices shared between the correct period, when the work was actually done
• start tracking the projects in the accounting software

I can tell you that they are not willing to do it in the beginning, but once you get them to understand your purpose, it will be smooth. Let me explain these actions with few examples:

First, the records in income statement – let's take the salary cost, which is the biggest input price for us. Usually we will find the salaries all combined onto one cost article along with the taxes. But in order to do business decisions you need to know if the salary is directly related to the projects or products you are doing, or is it indirect – like marketing. As they are manageable in different times and levels, they need to be separated.

Second – invoices for large fixed agreements might go out at the end of the project, or before the project. In order to know if you project was making loss and in which month the issues raised. We have introduced the aligned date in our accounting software, where costs are shared to period of value creation. This gives us month-to-month view on the dynamics of the project profitability.

As last example, get you incoming costs attached to the project. It is important that when you buy something specifically to support one or two projects, the costs land in direct costs related to the client account and the project.

Step two - start tracking the work time

As I said before – in IT and design world at least, your biggest asset is the team. It is also most expensive and least flexible asset you could have. Your decision are not immidiate and sometimes not even actionable. So you want to generate as much as possible data about that. You need to know exactly who and how much as worked on what.

There are some tools which help you with that. We are using a plugin in Jira called Tempo. It isn’t the most simple tool to use and for the beginners I’d recommend something like Toggle or even develop your own simple in-house web page to collect that information. We had our own internal system for years in the beginning. Reason why we changed was the integration to the software backlog – we wanted to integrate this not just to management accounting but also to software project planning and sprints.

Step three - integrate the data into one datastore

Now it gets technical and usually you need technical skills for this. Datasources I have integrated with have been Google sheets, Toggle work time recording tools, Jira Tempo work logs, POS systems, accounting software. Some of them require coding skills to transform the data, some I’ve been able to integrate with by just configuring pre-made service for it.

Process is called ETL (extract, transform and load) and if you cannot use tools like Stitch Data (https://www.stitchdata.com), you might want to hire somebody to do that, or hack together yourself. As we have the skills in-house, it was easy for us to start building these scripts.

Basic idea behind the process is that you will bring all relevant data to a database that you can later use for you analyses. We use Google BigQuery service to aggregate the data together.

Step four - start defining your KPIs

As now you have the data collection and also aggregated, you’ve enabled yourself a way to start measuring whatever you want. It is difficult to say what KPIs are relevant to specific business model and team set-up. We have defined company level financial KPIs and this is also easy place to start. In addition to revenue & EBITA, we also measure gross profit. That is the profit after direct costs and before indirect costs. That metric is relevant to product managers and important short-term indicator how we are doing. But what makes it good is that it can be affected also relatively quickly.

In the situations as we are today, we have added some KPIs to handle the cash flow. So now we also follow some data describes us cash runaway:

• cash on account
• along with length of a runaway
• net working capital - how much assets you actually have if you’d pay all your depts

And we also follow some KPI-s which give us an idea how much we use the power in our engine:

• utilization rate (UTR) - percentage of hours occupied in paid client projects
• revenue per hour - usually per teams but also a company as whole

Step five - automate the reports with some simple tool

We are using Google Cloud BigQuery database and Google Data Studio, there are probably similar ones. Some more expensive and more complicated. Why google works for us is that we can actually use Google’s other resources to data migration, automation etc.

Don’t buy most expensive tool and hope that it will do the job for you.
Be sure you know yourself or have somebody in the team who knows basics of SQL and data management.

Step fix - take it with you

The whole idea of building the data visualisation was to use when you need it. As I’ve been in mobile applications business, the team has built number of data dashboards relevant to many B2B services. So I’ve used that experience and designed most relevant numbers to my self in a way that I can access them from my mobile device.

What information I need on mobile? Mostly the one that helps me in short-term decisions. For example we’ve helped the bar to have a mobile daily report how last night went. So you can act quickly and revise the next day plans.

Summary

Having a good BI data isn’t simple and not cheap for sure. But if you can hack together a good IT product or service, you can do put together relevant data as well.

I’m happy to help any company to archive step 4 or 5. But you need to take your own steps. It is probably a topic that we all could write books about. If you have your own experience or I can help you out, let me know.