How a National Buy Group is Solving the Small-Retail Data Gap

Challenge:

Whatever business problems you’re dealing with right now, it will pay off tenfold if you use your data, not your gut, to solve them. The question is, how do you connect the dots between where you are now and what you must do to get to a solution.

That’s where most businesses need help. This is the story of how we helped a large membership organization connect the dots in scenario with layers of complication.

The company brought us in to help them solve two core challenges:

1. How to get and deliver more value from their own data.

2. How to convince more members that sharing their data would be in their best interest.  

Understand the Issues

Our client, a well-established organization with thousands of members, helps small retailers who sell national-brand appliances, electronics, and furniture compete against large retailers selling the same brands. A member could be a local mom-and-pop store who competes with the appliance section of Lowes or the electronics section in Best Buy.

By joining our client’s organization, members get access to some of the benefits that large retailers have.

Purchasing power, for example, is a major draw. With so many members and twice that number of independent stores throughout the country, our client gets volume discounts from manufacturers and passes them on to members.

But large retailers have the added advantage of tons of data from hundreds of stores and their ecommerce sites. That’s a critical difference.

Retail giants are savvy at using this data to make informed decisions like:

• What products to carry in which stores
• How to market products
• What prices to set
• Whom to market to
• What features customers are interested in most when buying the products

Our client recognized that smaller shops must begin to compete on the data front, and felt it was the right time to seek outside help.

If they could access member data across thousands of stores and share this kind of information, that would put them in a different league. But there were many obstacles in the way, such as:

• Most members are small store owners who don’t collect data.
• Some collect data through ecommerce sites our client built to help them drive traffic to their stores, but the data is not clean, the depth of information from any single source is lacking, and there aren’t enough sites to make an impact.
• Members collecting data don’t do it consistently or in a way consistent with other stores
• Many don’t know much about the world of data and are scared to share what they have.
• Some see data as an advantage, so why should they “give it away” to other members? They would need to be convinced that sharing their data is worthwhile.

Between those issues and trying to figure out how to mine their data to provide value, this was the backdrop against which we were called in to help.

Our client’s ultimate vision is to get to the point where members see their data as a value offering, and potential members join not just for the buying leverage, but also to get their hands on this valuable data.

Solution:

In preliminary discussions with key stakeholders, we agreed that our comprehensive COMPASS assessment and process for solving data problems would be the right path to coming up with a solution.

COMPASS evaluations consist of the following steps. Each assessment differs because it focuses on your specific business problems and where your particular company, technology, people, and processes need help.

• Gather input from stakeholders in your organization
• Perform a full inventory of your current data and technology stack
• Build a custom MAP that visually lays out where you are now and the recommended steps to take to get to your goal.

All COMPASS assessments put into action Onebridge’s MAP (Modern Analytics Platform) framework, which breaks down the entire journey an organization takes to go from having a data problem to resolution.

We took all our experiences and lessons learned through hundreds of data projects and created a scaffolding of what to consider/do to solve a problem involving data – that’s what MAP is. It figuratively separates the journey into three sections, strategy, execution, and enablement. Each section has checkpoints of all the factors you must consider along the way if you want your project to be successful.

Just like your car is assessed against a 15-point checklist when you get your oil changed, COMPASS offers a similar, but much more thorough, checklist of things to think through along your data journey. During a COMPASS assessment, we don’t just conduct a complete checkup. We also fix what else needs to be repaired – or at least recommend how to make such fixes if that’s what the client prefers.

Strategy  

We began this evaluation by conducting extensive interviews with different internal stakeholders to find out what was working with their data, what wasn’t working, and what they wished they could do.

We assessed their data governance, quality, data architecture, platforms, processes, and more. After talking to numerous team members, we were able to build a picture of how they trust their data, supply it to users, and analyze it.

For the most part, the organization had a very sound data strategy, but the assessment revealed they needed help with business value alignment. That means they were struggling to align their data to their business value, which was one of the primary reasons they engaged us.

We explained what they would need to do to develop that alignment and included recommendations in our evaluation.

Execution

In our discovery, we assessed the infrastructure and looked at our client’s tactical requirements for meeting their objectives. We used MAP to check many factors, and the results helped us build a roadmap and implementation timeline during this step. We agreed to identify what absolutely needed to be fixed and fix only those things. Other issues could be addressed later.

Data Profiling and Semantic Modeling Issues

Regarding the rest of their infrastructure, we found that our client was very skilled at pulling all their data into their BigQuery data warehouse.

But they were doing limited profiling, so we showed them how to do better profiling of the data within BigQuery itself.

One thing they couldn’t do with their existing reporting tool was modeling, a gap they needed to fill. Their existing reporting tool was very linear in that it only lets you visualize one set of data at a time.

We recommended they get a more powerful business intelligence (BI) tool that would enable modeling. We suggested Microsoft Power BI or Tableau, noting that Power BI would be the best choice given their purposes, but left it to our client to make that decision.

We also helped them understand the value of semantic modeling. Semantic modeling means when your data comes in raw, you have to regroup it into things that go together.

For example, if there are three places where you have page-view data, that data would need to be grouped together in a larger pool of marketing data. Then, within the marketing data pool, that data needs to be pulled together into website traffic data, which in turn needs to be pulled into another group.

Our client was bringing in their data raw, but not putting thought into the source of the data or grouping it.

When you pull in data with no context like that, the acronyms or shorthand used to describe the data might not be familiar to all users. That causes a lot of confusion and problems – as it did in this case.

Data isn’t as simple as something that gets entered in your organization in one place and then spit out somewhere else. It’s more than two end points. Along the way, lots of different people have to interact with that data.

As an example, you might end up with a few places in the database where you have data on how many people visited a page on your website, but there may be nothing communicating what was the source of that data. It could have been Google, the website analytics for a platform, or another web analytics platform.

You’d have to speak to someone who could explain the acronyms or tell you where the data was coming from, who owns it, and what is it used for. Until you clear up this confusion and do the grouping, you don’t know what to do with the data. This is what we helped them sort through.

Our client’s situation became more complicated because not only did their data team try to take that data and build report out of it, but they also had members outside the company who would want to get insights from the same data.  

If you didn’t properly identify data as it came in, it becomes even more difficult to provide data to members –that’s the semantic modeling part of data modeling.

Because of a lack of semantic modeling, anytime requests for data were made in our client’s organization, it took a very long time to build the data reports.

Demonstrating a Better Way

To show to our client what a difference semantic modeling makes, we did data dump of all their data over two weeks, performed a rough profile modeling of all the data, and then demonstrated how easy it is to instantly build a report on anything you want.

In front of their eyes, we quickly dragged and dropped the data they wanted into a report and produced a final report, right there and then. Our client instantly understood the value of profile modeling your data to make reports easier.

Recommendations

We were done with the checklist, but still needed to explain how our client could provide better insights with their data and how to incentivize customers to share their data.

We put those recommendations into an actionable roadmap of what they needed to do first to get moving on this path, the decisions they needed to make, and the order in which we’d recommend tackling these problems.

Leveraging our own consultants’ expertise and backgrounds in retail and marketing data, many of our recommendations focused on how to better present their data to their three different audiences: their members, the manufacturers, and themselves.

Here are some examples of the numerous recommendations we made.

• How to represent their data according to how retailers view their data over time (retail calendar) and consider buying cycles.
• How to put data in better context for retail cycles.
• How to compare data before, during, and after purchase to identify when a customer wanted something but couldn’t find it, found something but struggled to buy it, or abandoned the cart.
• What opportunities they’ve missed because people were looking for product they didn’t have. This data would be important because retailers would know what to buy more of to meet demand. The data would tell how customers shop for products versus how products are presented through ecommerce and in-store. For example, our client’s data was centered around the model numbers of the products, but consumers search for products based on features, the same way manufacturers do. Something so seemingly small led to missing the mark on how to present their data. This slight adjustment would make a big difference.
• How to build contextual data. We showed them how to represent their data according to how different retailers view their data over time. So they would see contextual data, like the missed opportunities discussed above. They could also see how to compare data before, during, and after a purchase, or when customers abandoned a cart, etc.
• How to present data to different audiences. To fix the issues mentioned in the last bullet, we provided a list of different ways our client could present their data to provide better insights to their customers, the manufacturers, and themselves. We provided examples to follow and showed the client how they could achieve this with the infrastructure they had in place.
• How to incentivize members to share data. We suggested our client tell members they would send customer satisfaction surveys on their behalf, provided the members would share the feedback data with our client. That data would be the foundation of a treasure chest of the kind of customer data our client’s organization needed so badly. Having consistent data for numerous stores would build the value that would attract new members.

Enablement

Enablement is crucial to success. You can put the best processes and technology in place, but if your end-users don’t know about it, understand it, or use it, it’s wasted money.

Our client is focused on enablement within their organization, and they know they must also enable all their members’ organizations. We did a user-group study to analyze their members’ needs and understand what they wanted.

To encourage adoption, we recommended:

• Data literacy campaigns
• An evangelist identification strategy to get their best members to use the new features and share with other members how great they are – that would help get other members on board.

Results:

In a short period of time, our COMPASS evaluation was used to assess our client’s infrastructure, processes, data, and use of resources. During the assessment, we either fixed the problems we identified or taught our client’s data team how to solve them.

We provided an overall solution, recommendations, and a prioritized roadmap that walks our client through the steps required to solve their problems.

In relation to their two core challenges (see Challenge section), we:

1. Explained how to get and deliver more value with their data.

2. Provided recommendations that will help their small-retailer members feel more comfortable about sharing their data and understand the value they’ll achieve if everyone contributes to the data pool.

In addition, we gave them critical guidance on how to approach marketing and retail data in ways consistent with the industry, leveraging our expertise in those areas.

Because of this COMPASS assessment, our client now has peace of mind that they are doing the right things and moving forward in a logical way, focusing on their most pressing pain points.

‍

"We engaged Onebridge to provide consulting services to help maximize our analytics offerings to deliver greater value to our customers.

Onebridge embraced best practices from a retail marketing analytics standpoint and looked at our data itself to call out the types of insights that could be available to our customer base.

The Onebridge team worked closely with all the relevant members of our team to truly understand our business, our needs, and our data.

At the end of the engagement, they not only provided us with fresh perspectives on producing decision-driving insights from our data, but also guidance on how to better approach managing and delivering our data to customers."

       -- CIO of leading national-buying organization‍‍

Impact on Client:

• Provided business value alignment
• Laid out strategy to improve data value and business intelligence
• Explained how to incentivize members to share data‍