Patch Plants harvests insight from a bumper crop of business data

Profile picture for user jtwentyman By Jessica Twentyman October 28, 2020
Summary:
At a time when interest in gardening is in full bloom, Snowflake is helping data-driven decision-making to flourish at the online plant company.

Image of a man holding some plant pots
(Image sourced via Patch Plants)

Growing plants, flowers or food can be good for the soul, especially in anxious times, as plenty of Britons discovered during lockdown. According to a May 2020 GlobalMarket survey, gardening was ranked the second most popular lockdown activity by 2,000 respondents, beaten only by watching TV and ahead of cooking, reading and exercising.

Analysts at GlobalData saw the findings as good news for UK garden centres, ahead of their reopening in mid-May - but the traditional, out-of-town garden centre experience isn't necessarily accessible to every keen grower, nor one that necessarily holds much appeal for them.

These are the customers that Patch Plants aims to target: a generally younger demographic, accustomed to shopping online and turning to the Internet for advice on seeding, planting, growing, harvesting. Its members, often apartment-dwellers, may not have a garden at all, just houseplants, some pots on a patio or maybe window boxes. Either way, four-year-old Patch is on a mission to inspire them to get growing, with its e-commerce ordering and delivery model, its online advice and active social media presence.

This business model has already produced an abundance of data, but until recently, Patch struggled to harvest real insight from it, explains the company's head of product Patrick Johnson.

One of the things I've spent a lot of time on is improving our online shopping experience, so we developed new features to try and improve that. For example, we implemented new navigation features on our website a few months ago, to help people find the plant or plants that they were looking for, ones that would be happy inside a flat or out on a small patio. And we're using web tracking software to see how people are engaging with that feature - but what we couldn't necessarily see is how that feature impacted other really important metrics for our business.

Bunches of data

This is a challenge that Patch has tackled using a cloud data platform from Snowflake. Data from the patchwork of software-as-a-service (SaaS) products that the company uses to run its business is drawn into the Snowflake platform where it can be combined, sliced and diced for analysis. This back-end technology stack includes Amplitude for web analytics; Ometria for customer marketing; and Zendesk for customer service and support. Segment is used to pull data from the website into Snowflake and Looker is used as its front-end reporting tool.

What Snowflake has enabled us to do is connect data from all these different systems and to understand what impact new website features have on our net promoter score, or on our returns rate, or on customer contacts into our customer service team.

In particular, this set-up has shed new light on supply chain issues, so that improvements can be made, says Johnson. That's important, he says, because Patch owns and operates its own supply chain, rather than relying on third parties, but more importantly, because different plants need different levels of care and attention in the warehouse and in transit, so that they arrive with a customer in peak condition. Says Johnson:

Plants can be difficult to store and transport, so as we expand, we need to keep a close eye on our supply chain. Until recently, we were just in London and Paris, but now we're expanding UK-wide, and that brings new pressures."

With Snowflake, we were able to build a dashboard that connects all of our data from the supply chain and combines it with customer data from Zendesk, so we can answer questions if a customer has an issue. What supply chain did that plant go through, where were problems seen, where are the pain points in the new supply chain that we've been setting up?

The Snowflake/Zendesk integration, he continues, is also enabling the BI team at Patch to start using natural language processing to analyse the customer contacts that come through to customer services via live chat and email.

This enables us to review all of the text that sits in customer service tickets and identify patterns. Is it a complaint? Is it to do with plant damage? Is it to do with packaging, or with the delivery service? We track those on a weekly basis to help identify where we have issues and address them.

Propagating data literacy at Patch

The Looker tools for front-end reporting from Snowflake, meanwhile, are helping the whole Patch management team to become more data-literate and make more data-driven decisions, he says.

Every week, the senior management team sits down to review trade and that meeting is based on a trade dashboard we have set up in Looker. We literally work through reports on every aspect of the business in the past week, checking how we performed. And actually, there are a lot of people in different parts of the business now who have become very good indeed at being able to create their own reports, freeing up the business intelligence team to work on other stuff, such as building new models to answer more complex questions, rather than straightforward reports to answer more typical business questions.