I’m certainly not the first one who faces obstacles in implementing changes in the organization from the bottom up and not the last one. I want to share some of my observations, conclusions and mistakes that I made along the way. Perhaps someone was in a similar situation, faced with similar obstacles and he managed to overcome them. It would be interesting to hear their experience.

The story is simplified as follows. There is a need for specialists in Data Science. There is some vision of top management about how it should be and it ignores people who are interested in this direction and are already employees of the company. Some of them study DS on their own. To speed up training in this direction and make it more relevant for the needs of the company, I decided to organize a group and enlist the support of architects and heads of departments. Here begins the most interesting in this story. Under the cut you’ll find the strategy and my conclusions.

Strategy: from erroneous to working

Everything looks much simpler if the idea to create some group arises in the heads of top managers. The group will always be created and there will be people in it. But if this idea arises at the level of developers, the prospect of its implementation becomes more complicated at times. But perhaps this is only my experience, from the product company and everything is opposite for outsourcing one. Against this background, the following understanding of the strategy emerged, which can be described as follows:

1. Find out the relevance of the initiative
2. Enlist the support of architects
3. Form a group (unofficial launch)
4. Discuss activity with heads of departments
5. Bring information to project managers
6. Official launch

After that, the initiative will be understood and supported by all. Ideally. Of course, this plan is not unidirectional: steps can be taken several times, if necessary. The purpose of them is to make the result and activity within this initiative the most relevant and safe for the company. The mistake here is that too many people are involved. In itself, this is necessary, but the order, in my opinion, should have been different, so that it worked. Now everything is stuck at the stage of discussions and the vision of the strategy has changed slightly. I’ll return to the strategy a little lower, but for now I want to clarify some steps.

1. Find out the relevance of the initiative

Here it is necessary to find out not only the number of people, but also their level of interest in the initiative. Roughly speaking, one must understand how many people support the initiative, and how much they are willing to spend on it their time. Later, the initiative will live at the expense of the second type of people. But even they should be interested in some exhaust from this activity. Therefore, even before this step, it is necessary to understand the relevance of the initiative for the company. It’s easy: talk to the CEO over a cup of coffee, if possible.

2. Enlist the support of architects

Architects are needed to adjust activity in the group regarding technology plans. Perhaps, there is no vision at all, because there are even no projects in this area. Nevertheless, the activity of the group can contribute to their appearance in the near future. It is best that architects know about the group and understand what it does to make timely use of the result. Or, on the contrary, give an idea for activity. In the second case, they can support the group when they request investments in the training project.

3. Form a group

At this stage, the final positions and the plan of the group’s activities should be discussed with the people concerned. How to work, on what, where to communicate, whom to involve, – and other issues should be discussed, to understand who is the backbone of the group at the initial stage. This step is necessary in order to come to the heads of departments. And, most importantly, it’s all brought to the attention of the architects, so that there are no misunderstandings.

4. Discuss activity with heads of departments

In some companies there is only one, in some more. In my company two, one for each of the directions. So, in order to provide them with information, it is necessary to operate with facts, not just considerations. At this step you’ll tell them that there are interested people, the architects agreed with the vision of the group and are, in a sense, mentoring it. The group itself builds a plan and operates outside of working hours. This is the most important point: the main activity of the company should not suffer. People involved in the initiative spend their time on it until the company approves the activities at a higher level.

5. Bring information to project managers

With the support of “Head of …”, you can already tell project managers that there is such activity and how it does not affect the daily duties of the people participating in it. It is necessary to remove the level of tension that always arises when something happens without the knowledge of managers. If this is not done, then very quickly there will be friction and questions on their part. During this discussion, there may be moments that need to be taken into account in the group’s activities.

Working strategy

The main mistake of the strategy above, in my opinion, is a large number of conversations and involved people. Involuntarily, this initiative is perceived as something global, as a separate project. Although, in fact, we are talking just about a club of interests. The idea is to unite people who already study the same thing in their free time and find synergy with a company that is only considering the possibility of moving in this direction without having a clear idea. Now I think that a more working strategy would be the following:

1. Find out the relevance of the activity
2. Formation of an informal group
3. Notify own managers

As the knowledge, experience of people in the group increase, and as the company develops the vision of Data Science, the relevance of other steps from the plan will come higher.

Findings for the Future

• What is KISS for you, not KISS for another person. What seems simple to you, can be difficult for another. The difference is greater, the greater the difference in the roles of two people. Even after it seemed that you reached the most simplified form of your idea, there are still different questions and unnecessary moments that can be thrown out and clarified by other people.
• You need to start very quickly and in the most simple form. I emphasize the “very, very simple” form. It means you have to sacrifice and throw away almost everything that seemed necessary your minimal version. If the topic is relevant, people will join and activity will be conducted.
• The number of discussions is directly proportional to the apparent size of the activity. Even if the activity is small, because of the number of people involved, it can swell up to a huge project, which will be an obstacle to its implementation.

P.S. It seems to me that this is somewhat reminiscent of launching a start-up, but within a company.