data leader
Blogs home Featured Image

Welcome to our Analytics Leadership Series of blogs, where we investigate key challenges for analytics leaders and how they can:

  • Effectively demonstrate a return on investment (ROI) from data and analytics
  • Make a practical move to an open-source analytical environment
  • Master stakeholder relationships for a strong company data culture
  • Implement a Best Practice approach

In this first blog of the Analytics Leadership Series, we look at what makes a great analytics leader and the core challenges they face.

The qualities and skills of a successful data science team leader

A great analytics leader will succeed in four key areas. With some flair, they will be able to:

Match the right technical skills to business requirements

It takes understanding of data science use cases and a change mindset to align the demands of data-driven business improvements to the technical capabilities of the team so that you achieve the required solution for the business.

Proactively identify opportunities for improvement

It’s one thing to respond to immediate business needs – for example addressing a fall in conversion rates. But a great analytics leader will also identify opportunities from experience and through deep domain knowledge to help improve business performance– for example learning hidden patterns in customer behaviour for personalising campaigns.

Create a culture of learning and development

A good analytics leader will manage hiring and team structuring. However, a great one will maintain high levels of job satisfaction amongst its data scientists and analysts by creating a culture of learning and development to constantly upskill their team.

Inspire and encourage innovation

In a role that touches multiple scientific disciplines, a great analytics leader will inspire innovation by encouraging team members to hypothesise and experiment.

Guide businesses through a data journey

It is the responsibility of an analytics leader to tease out the “unknown unknowns” in a business, e.g. understanding the importance of unconstrained demand and the appropriate KPIs that measure the success of demand forecast.

Core challenges facing today’s analytics leaders

As an organisation matures, the remit, challenges and priorities of the analytics leader will change. Over the course of this ever-evolving journey, the analytics leaders will face a series of core challenges.

Understanding WHY the business needs data

Pinpoint the business objectives. Understand and quantify what the organisation wants to improve on to achieve its business goals and work out which of these will benefit from using data.

Knowing WHAT to prioritise 

There are plenty of prioritisation frameworks available to help leaders, such as value / complexity matrix, and RICE (reach, impact, confidence and effort) that will help leaders quickly estimate the value of a project and a path for up-scaling.

Managing WHO to hire/work with to build capability

The analytics leader has several options here. You can either:

Hire directly: importantly in a competitive market, this approach is good for retaining knowledge in the business.

Contract staff: this is a more straightforward way of directly aligning human resources to a business case for well-defined, short-term projects.

Use service firms: this approach usually involves a high initial investment but is great for kick-starting the data journey.

Take a hybrid team route: this involves augmenting teams to balance in-house know-how with outside expertise, which also helps build in-house knowledge in the long term.

The challenge of HOW

From a practical point of view, a key challenge for analytics leaders is how to get things done. For example, how to de-risk data projects, how to measure success, how to tie a project’s outcome to its value to the business, and how to make the analytics process a core part of the business so that it becomes a sustained business as usual (BAU) element.

Fundamentally, a great analytics leader needs to be a partner to the business, with a value-based approach to driving initiatives. This is a leader who will be trusted to make decisions when tackling these tough core challenges.

Are there any aspects of your analytics journey that we can help with?

About the Author, Xinye Li, Head of Data Science

Heading up our Data Science team at Ascent, Xinye’s technical expertise compliments the team’s skills with vast data led solutions. With degrees in physics/astronomy and economics Xinye started his working career in the marketing effectiveness world at a consultancy. From there on he’s worked at both agency and in-house data teams across a myriad of industries. Enjoying the hands-on aspects of project delivery, he also likes getting his hands dirty and solving problems with structure and logic. He joins the Ascent team with innovative experience gained from half of his professional career based in consultancy and client-led senior data roles.

managed service
Blogs home Featured Image

In a recent webinar, we provided an overview of our Managed RStudio platform and demonstrated how modern technology platforms like RStudio gives you the ability to collect, store and analyse data at the right time and in the right format to make informed business decisions.

The Public Health Evidence & Intelligence team at Herts Country Council demonstrated how they have benefitted significantly from the Managed RStudio – enabling collaborative development, empowerment and productivity at a time when they needed it most. In turn, they have been able to scale their department.

Many of the questions from the webinar focused on the governance and security aspects of Managed RStudio. In this blog, we’ve taken all your questions and have for further clarity attached a document that can help with any further questions regarding architecture, data management and maintenance.

Many of the questions asked were aligned to the management of data in the platform from the process of working with data on local drives, user interfaces to the management of large datasets.

There are several methods of getting data in and out of the Managed RStudio. These methods will largely depend on the type and size of the data involved.

For data science teams to work productively and deliver effective results for the business, the starting point is with the data itself. Data that is accurate, relevant, complete, timely and consistent are the key criteria against which data quality needs to be measured. Good data quality needs disciplined data governance, thorough management of incoming data, accurate requirement gathering, strict regression testing for change management and careful design of data pipelines. This is over and above data quality control programmes for data delivery from the outside and within.

Can you please elaborate on getting data into and out of the Managed RStudio platform?

Working with small data sets (< 100Mb)

For smaller data sets, we recommend using RStudio Workbench’s upload feature directly from the IDE. To do this, you can simply click on ‘upload’ in the ‘file’ panel. From here you can select any type of file from either your local hard disk, or a mapped network drive. The file will be uploaded to the current directory. You can also upload compressed files (zip),. which are automatically decompressed on completion. This means that you can upload much more than the 100Mb limit.

Working with large data sets (>100Mb)

For larger data sets or real-time data, we recommend using an external service such as CloudSQL or BigQuery (GCP), Azure SQL Database or Amazon RDS. These can be directly interfaced using R packages such as bigrquery,  RMariaDB or RMySQL.

For consuming real-time data, we recommend using either Cloud Pub/Sub or Azure Service Bus to create a messaging queue for R or python to read these messages.

Sharing data between RStudio Pro/Workbench, connect and other users

Data can easily be shared via ‘Pins’, allowing data to be published to Connect and shared with other users, across Shiny apps and RStudio.

Getting data out of Managed RStudio

As with upload, there are several methods to export data from Managed RStudio. RStudio Connect allows the publishing on Shiny Apps, Flask, Dash and Markdown. It also allows the scheduling of e-mail reports. For one-off analytics jobs, RStudio also allows you to download files directly from the IDE.

The Managed Service also allows uploading to any cloud service such as Cloud storage buckets.

Package Management

R Packages are managed and maintained by RStudio Package Manager giving the user complete control of which versions are installed.

RStudio Package Manager also allows the user to ‘snapshot’ a particular set of packages on a specific day to ensure consistency.

The solution to disciplined data governance

Data that is accurate, relevant, complete, timely and consistent are the key criteria against which data quality needs to be measured. Good data quality needs disciplined data governance and thorough management of incoming data, accurate requirements gathering, strict regression testing for change management and careful design of data pipelines. This all leads to better decisions based on data analysis but also ensures compliance with regulation.

As a Product Manager at Mango, Matt is passionate about data and delivering products where data is key to driving insights and decisions. With over 20 years experience in data consulting and product delivery, Matt has worked across a variety of industries including Retail, Financial Services and Gaming to help companies use data and analytical platforms to drive growth and increase value.

Matt is a strong believer that the combined value of the data and analytics is the key to success of data solutions.

Fantastic data initiatives and where to find them
Blogs home Featured Image

It was great to be at Big Data London last week. Beyond the novelty of interacting with complete humans again (and not just their heads via Teams calls) I was also lucky enough to be asked to present on a topic that I’m passionate about: finding fantastic data initiatives that will generate business value and underpin a data-driven transformation.

In particular, I described the use of ‘data domains’ as a great mechanism to align business and data stakeholders and create an environment in which fantastic initiatives can be discovered. If you missed the talk, here is a quick summary of what was discussed.

The data opportunity

Data is the lifeblood of the digital age – delivering strategic value if allowed to course through the arteries of an organisation. It holds the potential to create smarter, leaner organisations able to survive and thrive in an increasingly competitive (and digital) market.

The Big Data LDN showcased the best and brightest technologies available to deliver on this vision, but the reality is that creating lasting data-driven value is not (just) about the tech. Building a data-driven organisation is more about culture than code, more about change than cloud, and more about strategy than software.

In my talk I focused on a topic that is pivotal to finding and prioritising great initiatives that will deliver lasting value for your business. I spoke about why this isn’t a trivial consideration, before describing our use of ‘data domains’ as a mechanism to create a focal point for discussion of specific data initiatives and a way to improve collaboration between business and data functions. But first, a thought about data literacy…

Becoming a data-literate business

Investing in data literacy within your organisation (building to a universal understanding of the language of data and the opportunity it represents) is an important way to connect both your technical and non-technical teams. By uniting people capable of building new data solutions with stakeholders in other departments, as well as enthusiasts who are intellectually curious about the potential of data, project leads and their IT teams, you begin to form cross-departmental working groups that offer more diversity of thought to address different business scenarios. These communities will come up with your ‘long list’ of initiatives – where ‘fantastic’ can be found.

Identifying fantastic initiatives that drive value

Whilst thinking up data initiatives seems easy, finding the exact right data initiatives is not. There are many reasons why this is difficult, but a major factor has to be the knowledge and experience gap that exists between business stakeholders and data practitioners.

Prioritising the right initiatives is essential to building confidence in the business around new ways of working. And it is incredibly easy to choose the wrong initiative, creating a capability that is disconnect from business behaviours, or something that works really well on a laptop but has no hope of being deployed into production or becoming part of BAU.

The role of data domains

As a team, we get to help customers discover the exact right capabilities to create. To achieve this, we use a range of mechanisms to create the right collaborative environment between data professionals and business stakeholders – in my talk I described the use of Data Domains as an important tool to enable fantastic initiatives to be discovered and prioritised.

We can think of data domains as themes that allow us to connect aspiration business objectives with specific data capabilities. Domains can be used to engage leaders and create a focal point for the discussion of specific initiatives, so that the right business outcomes can be identified and delivered.

Here’s 5 key domains that in my experience cater for the majority of customer scenarios and are highly observable in terms of their impact:

Using data domains takes us away from any language around the potential of data or AI, and instead focuses on the conversation on business aspirations. It also helps to prioritise and focus the conversation – of course, we all want to create more informed, engaging, intelligent, efficient, sustainable businesses but what is the highest priority? If we want to create a more intelligent business, then what are the most important decisions we’d love to get right every time? If we want a more efficient business, then which process would we love to automate?

A case in point. 

One of the best examples of the power of using domains I’ve seen was when working with a major insurance company.  This company had previously invested in data through the lens of a series of innovative initiatives focused on solution areas.  They had invested in ’an AI initiative’, ‘a blockchain initiative’, ‘a big data initiative’ etc.  None of these investments had delivered any value, and the leadership were running out of patience with their data programs.

When we collaborated with them to help them move forward, we looked at the use of data domains to create a more business-focused framework against which to guide investment. We determined that the ‘intelligent’ lens (iteratively enabling more effective decision making) was the best domain to start with – and one that would deliver the impact they were looking for.

We refocused the investment on decision improvement. We identified key decision-making processes across a range of business areas (from resource forecasting to customer call handling), turning them into more effective, data-led decisions.

Engaging stakeholders: building bridges with data domains

The example above was a great success because it created a consistent language across the business for data-led value generation and enabled business stakeholders to more quickly buy into the change required. Designing successful data initiatives using data domains can be an effective way to create a bridge between functional areas successfully aligning business outcomes and data potential. For example formulating what you are trying to do / become vs what you are trying to build.

Are you ‘ready’?

Identifying a fantastic initiative is the first part of the solution – but we need to consider an organisation’s readiness to undertake the initiative. Some tough questions to ask yourself include:

  • Are we ready to implement the change required to realise the value?
  • Do we have the right data available to succeed? How do we address this if not?
  • Do we have the right platform to support this initiative?
  • Do we understand our stakeholder needs well enough? Are we able to see the situation through the right frame(s) of reference?
  • Can we measure and report on our impact?

In summary: 4 step action plan

  • Build data literacy – Invest in building understanding of the data opportunity with non- technical audiences. Spark curiosity and enthusiasm. Connect and unite your technical and non-technical teams. Celebrate ideation.
  • Use data domains to define ‘fantastic’ – Zero in on the impact you are looking to create. What initiatives will deliver this impact? Where will you ‘see’ value – and in what timeframe?
  • Engage stakeholders and build bridges across the knowledge gap – use data domains to engage leaders and create the focal point for discussion of specific initiatives, allowing an idea to become structured and permitting objective evaluation and prioritisation of projects.
  • Develop organisational readiness – Think through critical success factors and objectively assess your readiness to take on the level of change you are proposing. Implement any required corrective action – or resize your ambition accordingly.

Data domains provide a great framework for discussing and evaluating high-impact initiatives with business stakeholders, guiding you towards the most ‘investable’ initiatives that deliver the most value to your business.

Rich Pugh is Mango’s co-founder and Chief Data Scientist. If you would like further discussion around identifying and prioritising the right data initiatives in your organisation, contact us and we’ll be happy to help.

 

BigDataLDN
Blogs home Featured Image

Big Data LDN (London), Mango Solutions’ big event takeaways: The ideal forum for sharing data expertise  

What better glimpse of the 2020’s data and analytics community than at the UK’s largest data event, Big Data LDN – and how great to be back! With 130 leading technology vendors and over 180 expert speakers, the London show on 23/24 September generated thousands of data and technology-led conversations around effective data initiatives, with real world use cases and panel debates. Everything from leadership, data culture and communities to the ethics of AI, data integration, data science, adoption of the data cloud, hyper automation, governance & sovereignty, and much more…

With a captive group of decision-makers, Mango Solutions conducted a Data Maturity floor survey of Big Data LDN participants. Respondents were asked to identify where their organisations are on their data journey, including the opportunities for data science strategy articulation, data communities, and what their biggest challenge is; with even the most data-advanced organisations modestly commenting in some areas that they had ‘room for improvement’. The fascinating results of this survey will be revealed shortly.

Conversations were approached with eagerness, around ‘harnessing data’, ‘getting more of it’, ‘data platforms’, ‘monetising data’, ‘kicking-off some AI’ – Big Data LDN being the first in-person analytics event since Covid-19 up-scaled our worlds of work and data consumption. From bright computer science graduates, to startlingly young data analysts, through to a surprising amount of MD and C-suite representation at the other end of the scale. Companies from Harrods to UBS, University of St Andrews – the best of the best of brands came out to share insight and learn – buzzwords aside – the most investable routes from potential to advantage.

A low-key event, but highly intelligent, Big Data LDN gave us a feel for the new ambitious, hybrid workers. Back to face-to-face networking, sizeable audience sessions, hot-desking in the cafe; there was a positive hungry-to-learn feel right through the event. Everyone, whoever they were, seemed on a mission to share intelligence and opinion on data and advanced analytics. No one job title or company outclassed the other – just a mutually supportive community of technology experts; less ego and more mutual respect.

There was the ability to attend a suitable meet-up with practical hands-on use case examples. Mango hosted a data science meet-up with guest speakers from The Gym Group and The Bank of England as well as our own Mango Consultant sharing best practice, which was well attended.

Chief Data Scientist at Mango Solutions, Rich Pugh’s talk on finding fantastic data initiatives was well-received by the audience. This focused on prioritising high impact data initiatives with buy-in from across the business. In particular how the role of data domains, supported by data and literacy, can create bridges between business ambition and data investment. In our experience, too often pitfalls exist on selecting the right data initiatives or even trying to answer the wrong question with data; so keeping a focus on achieving business goals and objectives is critical.

But as well as finding out about the latest advanced data services and technologies, this event was a great opportunity to step back from the day-to-day and be reminded of what needs to come first in any data strategy – achieving company goals, and measuring progress, in the form of clear KPIs which we can impact with data and analytics, before jumping in to make hasty data decisions. In this way, with a clear link between business goals and data, we can measure the return from data investments.

My main ‘don’t forget’ take-aways from the event:

  • It’s about aligning data initiatives with business aims. At Big Data LDN, there was consistent messaging around an ‘outcome-first mindset’. To make sure you start with business outcomes and objectives before considering the data, software and platform. It’s a valuable mantra to operate by, to make sure you always deliver the right data science, BI or software.
  • Focus on dis-benefits as well as benefits. As much as we focus on what things we should do to get value from data, we mustn’t lose focus on stopping doing the things we shouldn’t keep doing, or considering the opportunity and benefits lost when we overlook or don’t implement a particular data capability.
  • A robust data strategy must show the benefits for each stakeholder. Different people need different things from a data strategy. To gain buy-in and approval from all stakeholders, you must clearly show each of them the value they will get. In return, you will get not only their sign-off, but also, more valuably, their advocacy.

A parting word

We all learned many things from Big Data LDN 2021 – For me, I’ve learned our industry collaboration and community of data experts is unquestionable. We know remote working works, but you can’t beat the energy of this in-person event to confirm why you love doing what you do. For me, sharing our data science insight, and learning more about our peers’ technologies and services was truly rewarding.

Big Data Challenge – Our survey respondents were asked their no.1 post-Covid data priority, and we received a real mixture of enlightening and some more predictable responses. We will share these challenges with you and explain how we can support in our analytic leadership series.

As a relatively new member of the Consulting Team, working alongside Rich Pugh, Jon will be working with our clients and supporting them through their data journeys. He brings significant expertise to the team from his experience in data maturity, data & analytics, data visualisations and implementing strategies that inform successful business change and improvement, from his previous roles at Oracle, EY, HP and IBM.

If you would like to speak to a member of our consulting team about your data-driven journey, please email us.