Data Centricity is more than just having a data guy, it takes culture
“Without data you are just another person with an opinion” — W Edwards Deming
Remember the last time you were at a networking event where you overheard an entrepreneur or manager announce proudly just how data centric their company is? Unfortunately, being a data driven company takes a lot more than a badge, or even a cracking team of data gurus to crunch your numbers, it takes a wide reaching analysis culture to propagate throughout your entire company.
As a resident data scientist at busuu, I am always endeavouring to add this piece to the already world class culture of our team. In this article, I will list the steps that I have taken so far to achieve this and what the key successes are. I am privileged to be involved in projects within the marketing team, finance team, management team, product team and educational content team to name just a few. Working with them, I investigate and answer crucial questions like: Where is busuu currently? Where it is going? Towards what bearing should we steer the busuu boat to reach our ambitions? However, these insights have traditionally taken the form of a data product delivered by the data team, complete with an initial spec, adjoined Jira ticket, a set list of deliverables and granularity specific to the target audience and the hypothesis in question. As we scale into a larger company, this methodology becomes increasingly untenable due to the need for prioritisation towards more important projects. This hinders other investigations coming off the ground, and leads to bottlenecks for the data team when updating insight reporting, ensuring reuse, relevance of reports and the widest audience possible.
At busuu we have a very strong company culture that rewards curiosity and the growth of each person within the company. This is true not only in terms of their skills growth but also through having a front row ticket to the future world of startups and ed-tech that is unveiling before our very own eyes. I feel that adding more analytics and data centricity at a personal level is essential to further fuel the engine of curiosity that has propelled busuu into becoming the largest language learning social media platform on planet earth. Crucially, looking ahead, data centricity will bring us ever closer to our vision of empowering as many people as possible around the world to learn a language.
Attempting to promote this culture change and curiosity for data has been a major aim of mine at busuu so far. I have found that when you acquaint every member of staff, of all levels of technical ability, with the raw data points that affect the projects they are working on, a new branch of questions and excitement flourishes, based on what each person finds for themselves. This spark also promotes more collaboration from different viewpoints that are seldom achieved with the aggregated final reports that would otherwise be delivered by the data team. This goal not only has a positive effect on staff skill growth, collaboration and excitement but also allows the data team to widen its focus on other projects, such as growing analytics capacity by adding more datasets from inside and outside sources. This cultural centricity also enables the outputs of aggregation jobs and machine learning models, ultimately helping to improve the data products within the website and apps themselves.
Now, I will openly admit I have not fully achieved this goal yet, as staff, ideas and company priorities change over time, this aim itself is always in flux will never fully be complete. Nethertheless, I would like to share with you some of the key initiatives that I discovered so far to continuously promote and increase the data centric culture within busuu.
1 Accessibility & hitting the ground running
As well as curiosity, trust is another cornerstone of our company culture. To progress this further, I have worked to open up non-sensitive swathes of our raw data mart and coupled reporting software to all full time staff within the company. Every new starter is given access credentials to the necessary data in their first week, helping them immediately gain a sure footing into the structure of our existing analytics reporting. This allows everyone to then explore the current and past state of the entire company and find opportunities to help progress the company as early as possible in their time at busuu.
To make this level of accessibility possible and to maintain data protection and privacy for our lovely users, all personal and sensitive information has been hidden within this data mart. We achieved this by keeping sensitive data elsewhere or by using non-identifiable ids for individual users.
2 Automation
The great minds we have at this company shouldn’t be wasted on repetitive tasks such as rerunning current or historic reports or metrics. Not only do the industry standard Business Intelligence tools such as Tableau® and the new AWS Quicksight tool increase the speed of the process from question to answer, they are great for building reports that auto update and show a live feed of new data. Even though they might lose some panache, these reports can also be made as detailed as possible so that, when downloaded, they have the ability to completely replace companywide excel spreadsheets that waste time and effort to update and check. There is also a growing trend for other customer facing third party SAS services such as Zendesk® and an array of marketing CRM systems to provide their own viewports into user behavior; this can reduce the need and lead time for data pipelines to process the raw api feeds from these companies so that more time can be spent analysing instead of waiting for/preparing the data.
3 Celebrate the good news and the bad news
The only bad news is no news. The age old mantras that you learn more from a failure than a success and that, “the only strategy that is guaranteed to fail is not taking risks” are the main reasons behind AB testing and marketing control groups. However, when personal OKRs and the company KPIs that you are responsible for fall short of a target due to failure of these tests, it can be hard for teams and individuals to stay optimistic. Failed hypotheses should be celebrated within a data centric company culture the same way successes are because both represent new knowledge and increased company performance sooner or later. To some, this may appear counter-intuitive but if ignored insights are lost. Celebration of new knowledge — in various forms — is key to startup growth and can lead to crucial pivots at every level of the company to progress further towards future success.
4 Education
Having access to databases and data sources is one thing, but the reason that the data team had previously faced such bottlenecks was the lack of skills within the wider company to prepare and execute analysis of the data. This year I have been implementing “gimme my data” sessions to help people of all technical abilities and specialties analyse data using mainly raw SQL and Tableau®. Initial sessions went through query syntax, filtering and basic table joining and further sessions followed more of a surgery format, during which an analysis question was brought forward by any of the participants and the whole group worked together for an hour to solve it using the full array of tools available.
This program was met with great success; everyone felt that they were truly building new skills and had new tools available to make better decisions within their own roles and projects. Other subsequent effects of this were small but important in order to keep this initiative in place for the long term. These included a slack channel for everyone to help troubleshoot each other’s SQL syntax and data queries and the entire product team deciding to challenge themselves further by continuing their data centric journey through more thorough online courses. Even for myself as someone in the junior/mid level stage of my data science career, heading this process consolidated my own learning and tested my understanding as I shared this knowledge and we grew together as a team.
5 Focus
To many BI professionals in smaller companies, insight means keeping up to date with everything that is happening within the company the whole time. However, the reality is that as a company becomes larger and moves into an increasing number of revenue streams, product lines and specialisations, this again becomes impractical without change. As a data centric company culture has spread and more are staff members analysing a greater variety of raw data points for themselves, a lot of new opportunities for new projects within the business appear that had previously been missed.
Our management team does an excellent job to consolidate our efforts as a whole company by setting us challenging targets through some meaty top line company KPIs, that we all, as a company and as individuals can aim towards and be proud of when we achieve them. With a more ingrained data culture, I have seen a greater number of colleagues add to this with an increasingly diverse spectrum of personal KPI metrics that are not only more measurable than the personal targets used in the past, but are also specifically personalised to them and their role within the company. This measurement culture has been made possible due to the increased ability to monitor and drill down into them on their own to find opportunities within the data to achieve these targets and succeed.
— — -
A company wide data centric culture is not only a bold ambition but a hugely rewarding one too. It has so far helped to support stellar growth at busuu whilst increasing the pace of projects with quicker decision making and increased resources for further infrastructure improvements. As previously mentioned, this is a project with no defined start or end points and many more initiatives to go. Of course, busuu is not the first company to try to build a more data centric company, so comment below with your own data culture initiatives and let’s encourage each other to achieve this amazing goal.
I hope you enjoyed the post! And if this sounds great, you’d like to progress in your career, and you’ve got a love for learning languages, we’ve got lots of open roles in our team!
