Data Science Team: Value Creation and Stakeholders. The nitty-gritty.

Aim

• In this article we follow up previous article and go to the nitty-gritty.
• This could help to further understand the different dimensions of value creations of a Data Science team and how better leverage those.
• Why to read this? Perhaps you get ideas of how to improve your value creation.

Mapping Stockholders and value creation

From previous article we have the key stockholders and overall line of value we could further go into details for each of the cases. I believe this could be a good summary:

Value Creation from a Data Science Team and Stakeholders

Now we write a little bit more about all of those.

Value for customers

1. Problem solving. Data expertise and scientific and systematic approaches could render value to the business in alternative ways as initially considered. Sometimes there is no need for any data science like solution but an assessment and provide a solution which may be just a simple methodology used in a spreadsheet or similar.
2. Dashboard creation or augmenting. This could be a good solution, other times just adding in the top of a dashboard may be the right more valuable output.
3. Quantification of effects. This is sometimes difficult so it’s good to work out the best ways to do it. Perhaps some clever simulations, perhaps some effects measurement, or perhaps A/B testing.
4. Evaluation of data data sets. This is always something relevant, as many people only work with one a professional who has worked with many different datasets, or similar ones in different organizations, or just many others in the same organizations could have good ideas about how to combine and use. Maybe about similar cases within the company which could render economies of scale by granting permission to already created dashboards or similar.
5. Rationalisation and articulation of the problems. This is very common. The customer may not know how to articulate their need nor the art of the possible.

Value for business

1. Creation of code. Code could be intellectual property, or helps to support Business as Usual. Code is key.
2. Educational purposes
3. Showing the “art of the possible”. What is the potential thing we could do?. In many cases businesses need the input about things that are possible to do or not. Intrapreneurship (see Samit 2015) is something many companies seek when hiring Data Scientists. One thing is to see what is possible and another to actually try to do it.
4. Encouraging innovation, new perspectives and technologies. To have a Data Scientist around infuses people with curiosity and wants to ask and bring their problems to them. Maybe there is not a worthy solution but surely it is a good experience and a way to learn by doing. Closeness to the real source of value of the companies is great, as it allow for competitive advantages. This is key.
5. Refreshing and update the technology and data models. For instance Data Engineering has evolved dramatically in the last 10 years. New databases, new coding languages, a periodic refresh and evaluation of the data models and ways to administer this is relevant to keep the business well.
6. Diversification of offering. New products or different alternatives. AI and Data Science are in mind of many.
7. Evaluation of models and alternatives products. Sometimes it is good just to evaluate external tools or the creation of intervals. Creation of the company’s models and so on.
8. Model deployment and monitoring. This is one of the classics in terms of clearer value creation. Probably a key one. This is one of the most solid and clear ways to create value.
9. Scale: a larger team can support people of different specialties, holidays, sickness and as the different stages of the evolutions of the models.

Value for Employee

Value for Employee. Sometimes people forget about the employees. People will suggest the best model is to place a Data Scientist in a team, as this will immerse in the business case the practitioner. But who cares about the Data Scientist and her development?. Some people will mention best to have it working in local departments but link to and centralize Centre of Excellence or similar. Anyway, in many cases, businesses do not know what to do with Data Science less with a Data Scientist. This article aims to reflect on the best ways in this matter. The employee experience is important and another way to maximize value which many companies fail to unlock in the medium and long term.

1. Experience in the company and teams. This could help for the people to move internally to interesting new roles or similar.
2. Learning. In many cases Data Scientists are curious but also there is a need for adaptation to the business.
3. Datasets. For instance in Human Resources you have traditional Churn rational problems. Similar ones in marketing.
4. Technologies: databases, coding languages, cloud environments may be the most relevant.
5. Value creation. Implies the understanding of current models or analysis as the business the the technology readiness.
6. Techniques. Which may be the best techniques for the business and the available datasets. For instance if we have text from comments of customers maybe topic modelling or sentiment analysis may be relevant. If we have time series maybe times series forecast techniques are Auto Regressive models maybe relevant.
7. Development. In many cases it may be difficult to select the most interesting skills in the future. This is one of the places where good management could better support development. By providing the appropriate feedback, supporting the right amount of difficulties in task, and those to be oriented to a larger goal. Said that there is also generic blueprints in which you can see the differences between the different grades or seniority of the Data Scientist, such as https://www.gov.uk/guidance/data-scientist or D’Agostino & Malone (2019).
8. Soft skills. This is usually underestimated by Data Scientists and overestimated by business. To have the right balance will ensure promotions while keeping clarity about the value provided to the customers and business. This is something an experienced line manager can help a lot with.
9. Team dynamics. A team should provide much value that the sum of their parts and that maybe link to the dynamics of this. Agiles approach could be relevant. Good to have 101s, Plannings sessions, Retrospectives and Peer reviews.
10. Promotions. All the above should ensure the right valuation of the Data Scientist within the business. You want people to leave the team for being underpaid or for not having a good environment for growing. All of this will compound over time as people leaving the team may or may not come back in the future or introduce new people to those positions or not.
11. The career path for a colleague. As mentioned above many businesses do not know yet where to place Data Scientist, to promote, develop and recognise talent is key in any successful business and so to have a career path will help on that. It is not that difficult if you follow guidelines for other businesses and areas and articulate in the way business understand it.
12. A healthy team should be able to have a variety of skills and seniority. This will help for people to be mentored and to mentor, to learn from more senior people and be empowered in the right way. A common problem in starting a team is to have all the Data Scientists of the same level, perhaps under the idea of them to be mostly independent. That could be the right thing but in many cases in the future the team could benefit from a more tailored approach with different seniority levels and tasks assigned.

The Take Away

In this article it is presented in a systematic way the value creation and the stockholders of a Data Science team. I believe this could be useful to improve in the overall value proposition in the case of the reader.

Feel free to reach me to share any ideas or feedback.

The above only represent my personal views. Never the views of any of my employees.

References

• D’Agostino, M., Malone, K. (2019). The Care and Feeding of Data Scientists. O’Reilly Media, Inc.
• Samit, J. (2015). Disrupt yourself. Pan Macmillan.