Challenge
From family-driven to data-driven
AZP Multifamily is a family-owned supplier to multifamily properties, known for its customer service, and it set out to modernise on cloud technology so it could anticipate what customers need through difficult supply-chain conditions. It wanted more online engagement and a more personal experience for each customer, without losing the service it's known for.
Five things needed solving.
- An ERP that needed bringing up to date.
- No data platform and no analytics.
- Online engagement that could do more.
- No business intelligence strategy.
- A wish to use data in how AZP serves each customer.
Strategy
Azure around a Snowflake warehouse, starting with a prototype
The design settled on a hybrid: Azure components around a Snowflake data warehouse. The combination suited frequent, complex queries, deep analysis, machine learning, and personalising what each customer sees online. Machine learning and MLOps were part of the engagement.
The first step was a working prototype that ingested all of the analytical data.
Solution
Buy-again recommendations, and dashboards for each audience
Epicor Prophet 21 (P21), the ERP AZP had chosen, was optimised for performance and for AZP's specific needs.
Customer experience came first, because better engagement produces better data. When a customer logs in, the site recognises them, and a recommendation engine surfaces their past purchases and flags the items likely to need replenishing, based on how long each item lasts. The whole "buy again" function is personal to that customer. The visitor might even get a message saying, "Hey Bob, looks like your front office is running low on toilet paper," with a limited-time promotion attached.
Product availability sits front and centre, sourced from the ERP, which gives visibility into forecasting and replenishment. A customer managing several properties sees the data specific to each one.
AZP's data sat in P21 and a handful of marketing and sales systems. The warehouse brings it together with data from Google, the supply chain and elsewhere, and it is a working environment for data scientists. AZP also gained insights beyond the core business data, such as pool types, which inform personalised product offers, anticipatory reorder strategies and marketing planned around restocking schedules and seasonal trends.
Power BI reads the warehouse and turns it into dashboards built for each audience:
- Executive: cash conversion and sales performance against goals, for a KPI-minded audience.
- Sales: yearly, monthly and daily progress against goals, and the top-performing margins.
- Marketing: website trends, promotion use and shopping behaviour, consolidating Google Analytics and ERP data.
Results
Market share answered with a query instead of a spreadsheet
Marketing and sales kept asking the same question: what's AZP's share of a given region, and where is there room to grow? With property data in one warehouse, that's a query against stable records, where it used to be a spreadsheet assembled for each request.
Pipelines from real estate market data aggregators, alongside AZP's own regional data, mean national industry data can be stored until it's needed. That answers marketing and sales accurately and gives leadership what it needs when the company looks at a new market.
"Fantastic work. I like what they were able to accomplish with our website, particularly in the direction of data management reporting. Incomparable. So great. Great deliverables, really impressed."
Darren Rawson, President
Looking ahead
More data sources, and a new experience on the website
AZP found value in the data it already had, and keeps finding more the further it goes. It's enthusiastic about bringing in more data sources to create an entirely new experience on the AZP Multifamily website.