Business Intelligence

Proven way to roll out a dashboard and get users to actually use it

It might seem obvious, but one of the leading causes of newly-developed dashboards not being widely used or even looked at beyond the initial unveiling is the dashboards simply don’t show what people already care about. When developing a new dashboard, it can sometimes be tempting to go away into a ‘developer dungeon’ or keep the project worked on under the radar, until the big reveal. This drastically increases chances of not focusing on what the client or the target team needs – metrics like KPIs, OKRs etc – and instead implementing nice-to-haves.

When a dashboard does not answer questions people care about, or tries to do things they haven’t done before, there is a good chance it’s asking users to learn new habits in order to be used. This greatly reduces the chances of that dashboard actually being adopted – after all, everyone already has enough on their plates.

A better approach that could lead to a much higher degree of adoption involves focusing much more on what people already care about and do.

Don’t build new habits

Dashboards are much more likely to be used if they are incorporated into the habits and routines that people are required to perform throughout their day already. For example, if a team is required to compile and send Quarterly Progress Reviews, then implementing a dashboard of those Progress Reviews is a great target. Working with an existing routine gives you two things that drastically increase your chance of building something people will actively use:

  1. A well-defined scope to work with. This instantly allows you to ignore wasting time on building potential ’nice-to-have’s and continuous scope refinings.
  2. Concrete examples and reference points in the form of previous reports (in this case, previous Progress Reviews) that will greatly help with the implementation. They will inform you about the metrics, language and units that the team already uses, and are therefore more likely to quickly understand and use.

It’s easy to say ‘make a useful dashboard, as most dashboards are useless’. By focusing on integrating into the existing habits of a team, you can be sure they will understand and eager to adopt what you’re building together.

Business Intelligence

Understanding Your Customer Data

Common challenges in retail data

In large organisations, data is often siloed. This means that different teams within the workplace collect, store and utilise their own customer information. Often this is for good reason, because the data is optimised to suit the tools in use by each, individual team. 

It usually looks something like this.

With this type of data management, it’s impossible to get a high-level view of the customer from a total business perspective. 

With the data structure as it stands, it poses a number of challenges;

  • There is overlapping and redundant data stored by each, separate team. The cost of data storage to the organisation is high when data is siloed in this way.
  • The contents of each data silo differs slightly, due to each team setting their structures up differently, e.g. accounting might have different fields for storing customer names than the marketing team. Teams cannot easily determine whose data is most up-to-date and accurate relating to the same customer.
  • When teams inevitably need information from each other,  manual reports are prepared and shared, which quickly go out-of-date. Generating reports is time-consuming, and reduces the teams capacity to do other, more productive things. It also becomes impossible to tell if the decisions made based on the reports provided are using the most up-to-date data.

Lastly, but most importantly, no one team is looking at the same customer data. Executive teams are unable to make well-informed decisions, because the data they are dealing with is already out-of-date by the time they are reading it.

Get ready to know your customer

To solve these problems, three key strategies have to be employed.

1. Data integration

All the data needs to be in one place. Every first-party customer interaction, on all channels, needs to be brought together.

This means mobile, back-end, web, in-store, payment services, help desks and CRM interactions are all stored together and identifying the customer at every point of interaction.

Importantly, data integrations should also allow data to flow the other way as well – from the centralised data store to the tools the teams use.

2. Data governance 

Data needs be reliable, clean and uncorrupted to be of value. Governance is important because it ensures that once data is extracted from the application that collected it, the rules and details of the data are known and respected by all other applications, and by teams that use it.

Good data governance makes data easy to use and share.

3. Audience management

Data needs to actionable once it is collected. AM builds profiles out of the ingested data. It allows specific information about preferences and habits to be drawn from the customer profile. Things like brand loyalties, favourite categories and online log-ins per week should all be accessible to be easily acted upon by interested teams.

Customer Data Platforms 

A Customer Data Platform (CDP) is the gold standard tool for integrating, governing and managing data.

With a CDP, the data becomes consistent across all teams, and is of the highest quality, Audience management metrics are driven by complete and accurate data. 

The consequence of managing data in this way is that each interaction with your customer can be personalised based on accurate and up-to-date insight. 

It looks something like this:

Source: TreasureData.

Advanced capabilities unlocked by CDP

Providing unified customer data, enables advanced business capabilities, including:

Improve your user’s experience

Create magic online experiences that encourage customers to stay longer and buy more, backed by accurate feedback from user interactions.

Personalise product recommendations

Know your customer well enough to provide recommendations of your products and services, in line with their needs.


Know how your customers interact with your merchandise and plan your layouts and visual displays accordingly.

Customer loyalty and lifetime value

Know exactly what your customer is worth to your business, how loyal they are to your products and services and what makes them stay.

Online to offline

Follow your customers from online to offline engagement with your business. Know what each customer wants to do online and what they do in-store by connecting their experiences.

A solid data backbone can transform the ability to see deeply into a customer’s wants and needs. It can create magic experiences, where customer’s feelings and desires are easily interpreted and delivered upon.

CDP case studies

All case studies are provided by Deloitte.

Product Recommendation — E-commerce platform, EMEA

A global e-commerce platform wanted to expand its product recommendation capabilities across its many product verticals and [improve their data strategy]. The retailer spent 12-18 months aggregating all customer data across verticals to create a single view of customers’ browsing and purchasing patterns. The retailer ran A/B tests on the customer data and developed an algorithm to drive continuous improvement in recommendation accuracy. This resulted in a 500% increase in sales conversion in some of the retailer’s product lines

Customer Loyalty and LTV — Multinational supermarket, EMEA 

A supermarket with an established loyalty card program wanted to boost its effectiveness in driving long-term customer LTV. To improve its maturity from mature to leading, the retailer incorporated additional metrics into its LTV calculation beyond just spend (e.g., number of categories shopped, number of channels shopped, frequency of visits). The retailer also revamped targeted promotional offerings to personalise every message and offer sent to each customer. Success was measured by the ability to move customers up the LTV curve. By taking this highly personalised approach to customer loyalty, the supermarket increased its bottom line profitability by one percent

Online to offline — Multichannel furniture store, US 

A national furniture retailer found that, while its website attracted a lot of traffic, the majority of its sales happened in-store. The retailer wanted to improve [its data capability] to increase footfall in-store and improve sales conversion rates. The retailer developed a quick response (QR) code model to track customers who move from online browsing to in-store shopping. The QR code offered a prize, which was only redeemable in- store. Customers were directed to the nearest store using location data from when they first scanned the QR code. When the customer then rescanned the QR code in-store, sales assistants could view what they browsed online and show them similar in-store furniture, thereby improving the customer experience. 

Business Intelligence

Maximising the Potential of Product Data

Getting it together

Just like with customer data, siloed product data can cause major headaches. Tracking products across franchises, branches, international brands and warehouses is extremely challenging, even when operations are well-organised.

Product data is essential to operations. When the data is collected and stored individually across each arm of the business (siloed), reporting has to be done by each branch manually and then manually aggregated. Productivity suffers and decisions made based on this data are out of date by the time they are implemented. 

The structure usually looks something like this.

Get ready to see you inventory clearly

To solve this problem, three key strategies have to be employed.

1. Data integration

All the data needs to be in one place. Every store inventory, warehouse inventory and online product list needs to be visible in one place. 

This means pipelines need to be built to ingest data from each individual source to funnel it to one centralised data store (e.g. data warehouse or data lake). This is a prerequisite to accessing a holistic view of the state of product within the business. 

2. Data governance

Data needs be reliable, clean and uncorrupted to be of value. Governance is important because it ensures that once data is extracted from the source that provided it, the rules and details of the data are enforced at the time of ingestion into the centralised data store.

Beyond that, a Data Catalogue is essential to be created and kept up-to-date throughout the entire process. It allows self-service across teams, and reduces the need for back-and-forth on data definitions.

Good data governance makes data easy to use and share.

3. Modernising business intelligence (BI)

Business intelligence has always been an essential part of any modern business. Usually, when data is siloed, BI tends to also be siloed and only applicable to the teams that source the data. 

Changing the approach to a comprehensive view of product and inventory is difficult, when each team has their own approach. 

A solid data backbone that allows quick and robust integrations plus storage, allowing BI to be readily accessible across teams. It breaks down barriers between teams, allowing for better, more effective insight and self-service. Teams no longer have to wait for each other for manual reporting or question whether the data is up-to-date and valid. 

Achieving advance capabilities

Providing unified product data, enables advanced business capabilities, including: 

Optimise product movement 

A high level view of all stock-in-place gives a clear picture of opportunities for the movement and management of stock. Get your products where you need them most with ease.

Real-time fulfilment and inventory dashboards

Know the true state of all your product holdings across multiple sites at all times. A centralised data store can have a dashboard layered on top, with data updated as quickly as data integrations allow.

Assess stock levels and seasonal readiness

Visualise your historical sales data across the whole business and plan for seasonal sales, promotions and special stock requirements. See historical trends and plan accordingly.

Predict and prepare for demand

Tracking the stock flow over time, allows for a historical picture of key points of demand. By employing machine learning algorithms, seasonal and non-seasonal trends and patterns can identify opportunities automatically. This means you can put stock where it’s likely to be needed, when it’s likely to be needed.