In the retail world, data is king.
As Forbes note in an article addressing five growth strategies Consumer Goods (CG) companies can use today, “Gone are the days of making a decision based on a gut feeling.”
There are three main reasons why this is the case:
- The retail industry has become ferociously competitive with manufacturers now facing online retail giants, local challengers and private brands on top of their all-time competitors (CB Insights, 2018).
- Changes in shoppers’ tastes and a decrease in millennials’ purchasing power (McKinsey revealed they earned 9 percent less than Gen Xers used to at the same age) have disrupted the rules of brand loyalty, urging CG companies to reinvent their offering and brand strategy.
- The amount of data available to businesses today is tremendous. IDC showed that the global datasphere amounted to 33 Zettabytes in 2018, a number which will sky-rocket to 175 by 2025. That’s 33 billion Gigabytes of data processed last year, and another 142 coming our way within the next six years. At this speed and level of access, overlooking the hard evidence and opportunities that this data mine offers businesses would be a foolish risk.
Yet, McKinsey shows that while the benefits of business intelligence and analytics in the retail industry are many, the sector captures only 30 to 40 percent of the value they provide (2016). Furthermore, a study conducted by BCG with Google found that using AI and analytics at scale to measure retail execution performance is one of the must-haves for manufacturers to increase sales, but 69 percent of them are still at the very beginning of their AI journey. It is not due to lack of awareness; BCG reveals that most majors understand the value these new tools can bring them. However, they experience challenges scaling up their new apps such as:
- Insufficient robustness to deploy their solutions globally,
- Organizational and technical difficulties impacting user adoption and integration within existing systems,
- Data quality and framework issues to drive accurate and informed business decisions.
What’s the use of processing tons of data if you can’t trust or combine it to detect unforeseen trends and identify retail execution performance gaps?
A few months ago, we chose to embark on a journey to enhance our image recognition (IR) solution’s capabilities by designing a new portal to answer our clients’ most urgent retail execution analytics challenges. I interviewed the people at the origin of this project.
Here’s how Peter, Tamás and Fred made retail execution analytics part of Planorama’s IR solution.
Oh, and if you don't have time to read the entire interview, you can still rush to the section that interests you the most right below:
- Kicking off the retail execution analytics revolution at Planorama
- Helping our clients capture the best out of retail execution analytics
- What's next in the roadmap?
How did it all start?
Fred: I think – like most startups, we started building some tools directly based on our clients’ needs. At some point, we looked back at what we’d developed and realized we had lost track of our roadmap a bit. So, it was important for us to get back to the basics of what the product is and how it needs to evolve to deliver the best to our clients.
Peter: Our web app had had quite a few makeovers. But, as Fred was saying, we used to design and upgrade this app as our clients’ needs evolved. Back then, our service intervened in the second phase of the retail execution value chain, between data collection and retail execution analytics, with our image recognition solution turning photos into shelf data and processing this data to produce real-time KPI reports. We hadn’t fully measured how our ability to use business intelligence tools to crunch and analyze shelf data into enhanced retail execution insights was rare in the industry. Very quickly though, we realized that clients did not have the necessary resources to do that job internally. Consequently, they could not properly analyze the shelf data they collected from the stores and were losing part of the value of our solution, as a result. That realization prompted us to do a more thorough investigation about what our client users expected to see on their screens. We needed to conduct a proper user discovery. That’s when the idea of Portal was born.
What did we want to achieve with this new Portal?
Fred: That was the question we were asking ourselves. We continuously get valuable feedback from our clients, but we needed to take a step back to understand the users behind our products, what they want from the data, etc. That was the whole purpose of this study. It’s not the same thing as asking a UI designer to work on a new page mockup to modernize an interface. It’s about figuring out who are our users, what they are looking for, what they are looking at, why, what are their problems, and so on. We needed these answers for our solution to solve our users’ problems more efficiently.
What was your role in this project?
Tamás: As Product Manager for our old platform, Peter’s role was key in helping us understand the legacy and the history of the tool. I was the Project Lead for this study.
Fred: I wasn’t in the company at that time. (laughs)
So, how did you get involved in the project, Fred?
Fred: When I arrived at Planorama in 2018, the discovery phase had already been done. So, we had a final report with a 360-degree view of our personas - who they are, how they access the solution, their problem, as well as a UI study and recommendations. I got involved in the next phase when we used the audit to lay out a development plan. We decided to extract the most critical blocks we needed to deliver as an MVP and to draw the user journey based on that.
Peter: In other words, we prioritized our developments to deliver the highest value to our clients right away. We looked at their most urgent needs and at the impact our developments would have on our different personas and decided on a plan.
Fred: We did that? (laughs)
Peter: We did! At least this is how I remember it. (laughs)
How did we build our Portal for retail execution analytics?
Tamás: We wanted to be able to integrate with a new retail business intelligence (BI) tool to help our clients analyze the shelf data they get from our solution. Selecting the right tool was quite a ride. We’d had our previous one for a couple of years. We’d chosen it at the beginning of retail business intelligence and analytics. It was actually one of the first usable BI platforms in the market - I think PowerBI was not even released yet. But as time passed, we realized that it didn’t fit our clients’ needs anymore. Users wanted more flexibility and speed, so we benchmarked four other analytics tools to meet that need.
Tamás: In the end, we chose Microsoft PowerBI for its flexibility, speed, and cost, mostly. Funny enough, having the possibility to embed the tool’s analytics within our product was not a part of the criteria at first. We learned from a lot of clients that providing a single web application was much more convenient and user-friendly to them later only. Before, we used hyperlinks to help them navigate from one app to the other, but users often had authentication issues or trouble finding the hyperlinks, for instance. At some point, we had also thought about building our own retail execution analytics before embedding PowerBI.
Fred: Yes. But, doing so didn’t make sense because there were already many powerful retail business intelligence tools with extensive capabilities. So, wasting time developing a product which was already available and known by our users was not an option for us.
Tamás: Exactly. From the moment we learned about PowerBI’s embedding capability, it was clear to us that we needed to integrate with the tool to build on its features and provide our clients with the best way to increase performances in retail. It’s a tool that’s evolving fast so we can benefit from that.
Fred: And since we wanted to offer our clients a one-stop shop application from which they could see everything, PowerBI was the best choice. Plus, it’s one of the most trusted and used solutions in the industry, so partnering with Microsoft helped us deliver retail execution results in a language our clients understood. That aspect was essential for us because it meant our clients already knew how to use our analytics, how to customize the tool, how to exchange data with us, etc. That’s one of the main reasons why we chose PowerBI.
What is “Row Level Security” and which role did it play in the project?
Peter: Row Level Security is a must-have for our clients to limit what each user can see in the retail execution analytics. They don’t want unauthorized people to see some data which they are not supposed to see. That means they need a way to restrict access to certain types of data.
Tamás: The way it works in Planorama Portal is that we will authenticate the user and, based on their credentials, identify which level of access to give them in the user hierarchy table.
Fred: It’s a key element of Portal. We receive a lot of data from our clients, so we need to secure them. Our clients trust us to build an application where the security and access rights of their users are respected. So, we had to aggregate all these data in a way that every person could access the information he/she needed and was authorized to see (for the sales rep in the field to be able to see his/her performance, his/her team lead to see his/her team’s, for the country manager to see his/her country team’s, etc.).
Tamás: I also see it as an enabler for users. For instance, sales reps have more and more appetite for business intelligence and analytics in retail because they want to see a detailed scorecard of their performance, something specific to their stores. So, they want more information than what we provide in our live reports, which they receive while in-store. With Row Level Security, manufacturers can start empowering field forces with retail execution analytics by giving them partial access to Planorama Portal.
What are the benefits of the new Planorama Portal for our clients?
Tamás: The key benefit now is the retail execution analytics that they get from our reports and our integration with PowerBI. If we look at our roadmap, there is a huge appetite for a feature we’ll make available soon where users can save their store checks selection and share it with different people within the organization. The other benefit is around master data management, thanks to which users can edit the master data in a SaaS manner.
Fred: What it means is that, if during a store visit our solution can’t recognize a product because we don’t have its picture or any information about it in our system, we can automatically generate an alert in Portal to notify our client about the situation. They can then enrich their database much faster by simply attributing the corresponding product information to the photo of the SKU directly in the web application.
Tamás: Eighty percent of the time, the last question we hear after we build BI dashboards for our clients is: “How can I export the shelf data if I want to make a calculation that is not already available in Portal?” or “How do I export the pictures of store visits to insert them in my PowerPoint presentation?” From a product perspective, our goal is to cover such requests with the solution itself so that users don’t need to export data, pictures, do screenshots, or whatever. Portal will evolve to respond to these general needs.
Fred: We also see new trends around retail execution analytics and business intelligence. In the past years, Excel was like the standard for any data management in the retail industry. Everybody still knows Excel as the way we work with, analyze, and share data. But there is a shift to BI tools, which is very interesting because they are very powerful and allow businesses to make in-depth analyses of pretty much anything. The trick is that all these data take a lot of time for analysis. What we envision at Planorama is to bring Artificial Intelligence in this process. We want to be able to adapt our model to exploit retail execution data and automatically fuel our clients’ dashboards with trend detection, low signals, specific alerts, etc. That’s the direction we want to take. So, data will still be there, data will still be accessible, but instead of pointing out a problem, we’ll try to figure out why that problem exists and what our clients can do to solve it as fast and as efficiently as possible.
And again, this will be done based on clients' inputs, right?
Fred: Correct. We need our clients’ help to develop standard mechanisms. For instance, standard statistics about specific data clusters to understand what the levers of one KPI going down might be. Since we use deep learning, we need to figure out which metrics or tags our users look at to analyze their data, where and why they do so, etc. to be able to do predictive analytics, i.e. to teach the system to surface meaningful predictions and recommendations. Right now, we can try to model it, but we expect some users would be more interested in specific categories, or specific stores, regions, and so on. So, once we know what they are looking at and the rationale behind it, we can adapt our model.
Peter: And if we look even beyond retail execution analytics, what we want to achieve is for Portal to become a self-service application. So basically, a safe space that clients can access to review their execution performance, look at some business analytics, create and edit master data, etc. In the long run, we’d like users to be able to just log into the web app, easily set up their account, quickly configure their solution, and start using the service a few weeks later without Planorama getting involved in the process.
How much time does it take to set up Portal now?
Fred: It’s set up with the solution. So, once their account is configured, clients have access to everything.
Tamás: And if they have special requests, we can build custom dashboards, but it may take more time to define their scope, their content, and how they will work.
Fred: But that’s the power of using a standard solution. In the end, we integrate our clients’ shelf data in standard PowerBI dashboards available in our Portal so that our users can instantly see the value of the data we create with our Image Recognition solution. Then, we can go even further and adapt any dashboard or any visualization or insights to our users’ needs. We can even go beyond that and integrate their data with ours to help them combine their shelf KPIs with their sales information, for example.
Do you want to know more? Get in touch and ask for your demo now.