- Snowflake Analytics offers analytics expertise and deep experience with the open source Snowplow Insights system
- Oneflare now runs a centralised customer data platform using Snowplow Insights
- Customer data from all parts of the business is now standardised, integrated and accurate
- “Snowplow is at the heart of our decision making: all our customer data infrastructure, all our dashboards and AB testing – it’s everywhere. We picked the right tool to grow with us.”
- 87% growth in organic traffic and 110% growth in job listings (January 2018-January 2019)
Oneflare is an Australian online marketplace for home services connecting more than 150,000 qualified businesses across Australia with 1.8 million customers.
Customers can submit a job request free and then receive up to three competitive quotes from nearby businesses. They can then compare quotes, access profiles and read reviews to help them choose the right expert for their job.
Oneflare covers 300+ service categories from plumbers and electricians to pet groomers and interior designers who buy credits for quoting on relevant job leads. Approximately 15,000 new jobs are posted each week.
The challenge: driving growth by unlocking the truth about user experience (when data is in separate silos)
After five years of steady growth Oneflare accelerated its aspirations for a bigger share of Australia’s $130 billion home improvement market with a $15 million investment from Fairfax subsidiary Domain Group in May 2016.
Nathan Scully, who now leads the Platforms team at Oneflare, joined the business as a senior analytics manager two months before the investment was announced. He remembers a very strong push for rapid growth in user numbers and revenue. Both raised a big question: which customer-focused innovations would deliver the best results?
To that point, Oneflare was managing data in multiple silos throughout the business. There wasn’t a single source of truth:
“We have multiple inputs through our apps and four different branded web domains all at the same time and we hit a point where our existing data infrastructure couldn’t answer the questions we were asking,” Nathan explains. “Most of that was due to attribution of end-to-end and user journeys.”
“We also needed solid information across different systems. We wanted one solid piece of data collection infrastructure that we could pull in – and we wanted to improve our overall data governance.”
One big bugbear was what Nathan describes as ‘the Rails monolith’, inside which data was modelled with an engineering focus, rather than analytics.
For example, there wasn’t a way to capture changes to a business, so without an events log of changes over time Oneflare was potentially missing out on opportunities to tweak user experience to help its customers grow their businesses.
“We knew we wanted to do change capture events but we didn’t want to burden our Rails infrastructure with data and analytic capability,” notes Nathan. “We started looking for a way to get that change capture; and the other big thing we needed to solve was call log data. Soon we decided we’d buy those capabilities in a new system.”
In December 2015 Oneflare engaged Snowflake Analytics to help design a new analytics system, attracted by Snowflake’s offering of Snowplow Insights combined with expertise in open source analytics.
“From the start of our analytics journey we partnered with Snowflake Analytics to ensure we got data right at Oneflare,” says Nathan Scully. “Snowplow Insights enabled Oneflare to iterate quickly alongside our developing product, allowing the insights team to validate hypotheses and prove out product decisions that improved our core metrics.”
Another key reason Oneflare chose Snowflake Analytics is that the expanding business. Either leave it and attribute it to Oneflare or shall I reword it and it can be attributed to me
wanted overall ownership of its analytics system, says Nathan:
“We want all the data on our side, with full control. So if we need to change something we can open up the tools and fix it ourselves.”
“We also want guidance and Snowflake offers a good blend of the two. Its team is focused on technical outcomes to help us be successful and that resonated well with us.”
The solution: Snowflake Analytics supports customer-focused innovation
(1) Smarter data management across multiple domains
The first way Snowflake Analytics helped Oneflare was to instil best practices across the business for how data is handled.
Next, Snowflake guided Nathan and his team on modelling all customer data points from different departments and websites so they can be managed in a single data platform with Snowplow Insights.
Business-wide metrics are set by Oneflare’s product board, which is a governance group made up of the product team, head of engineering and the C-suite:
“So, there’s an oversight of general business health and then each squad has its own remit and metrics,” notes Nathan. “The biggest thing is, as we’ve acquired more businesses, we need to stitch together more and more data sources. We’re focusing on multiple consumer journeys and unifying our data infrastructure underneath with Snowplow as the portal.”
Nathan notes some advantages of using Snowplow for a unified pipeline include the ability to use analytics tools across other domains, giving Oneflare full stability of data collection when the asset isn’t owned:Nathan ScullyNathan notes some advantages of using Snowplow for a unified pipeline include the ability to use analytics tools across other domains, giving Oneflare full stability of data collection when the asset isn’t owned:
“The guys at Snowflake helped us set up our GTM containers in Snowplow using a lot of data variables, which allow us to do domain inference. We can use these containers across non-owned assets so the owners can pull them into their side or we give them a widget for our services inside their site – we love that flexibility.”
(2) Central core event pipeline
With robust definitions in place, Oneflare has a consistent and standardised approach to producing and analysing its customer data:
“We use Snowplow as the core event pipeline for all data collection across Oneflare, from apps to website services,” says Nathan. “Everything gets transformed into a defined event then goes through our enriching and shredding process to end up in our Redshift warehouse, where we build all of our ETL pipeline.”
Customer behaviours are tracked at every point, including button clicks, form interactions (and errors) and prediction events online, as well as activities in offline services, such as calls.
(3) Experience testing for businesses improves conversion
Businesses that sign up with Oneflare know that getting the right leads for services they offer, in the areas they can feasibly service, are essential to winning work.
Oneflare runs frequent A/B and other multi-variant tests through events in Snowplow to help it refine and improve the user experience, including ensuring every element is labelled and tracked correctly.
Tests in the jobs notification process, for example, help improve matching of the right businesses with valuable customers.
And attribution modeling via Snowplow helps Oneflare optimise content to improve search rankings, which Nathan reports has driven 87% growth for organic traffic and 110% growth in job listings year on year (January 2018-january 2019).
(4) Tracking business to consumer events improves retention
Likewise, Oneflare’s refinement of its refunds process helps retain businesses. Oneflare is committed to refunding credits if a quote or the business’s profile hasn’t been seen by a customer within 48 hours. By tracking all processes on the page, including profile reveals, clicks, timings and contacts Oneflare has an accurate record every interaction or inaction.
“We want to quickly refund tradespeople and other businesses who bid for a job but the customer didn’t respond,” Nathan explains. “It’s important we track all those events properly, otherwise the refund process fails. We also have timings in the system to make sure refunds are managed within 48 hours.”
(5) Outbound leads are accurately costed
Oneflare uses a Snowplow collector that tails a log on our call centre server and sends Snowplow events out for every single call event. This system is used by our sales and support staff for inbound and outbound calls, across multiple brands.
According to Nathan, setting up that feature was just one week’s investment in development time. The Snowplow SDK was easy to integrate, very functional and simple by design. And it hasn’t fallen over in two years of use.
“One of the amazing things that Snowplow collector has done is we can now do outbound lead attributions when we call the mobile number in a booking. That number has a booking, event or job in our system allowing us to stitch that entire journey together,” says Nathan.
Oneflare can now track all the costs-per-lead for outbound calls, including the call itself and sales team call time.
(6) Self-serve insights systems improve efficiency
Interestingly, investment in automated self-serve recording capabilities allowed Oneflare to decrease the number of people assigned to the insights team from seven to three, including one data engineer.
Insights specialists are now assigned directly to product managers, so they can focus on more specific challenges. Marketing, product managers and executives now all self-serve their data.
Oneflare’s conscious investment in data, insights and infrastructure is part of its product-first approach to engineering, rather than simply adding more analysts.
Whenever the business hires insights people, it looks for a blend of data science and insights with general business skills:
“Our insights people are very product-focused,” adds Nathan. “We transitioned from scaling up the analytics team a few years ago, to scaling our infrastructure and engineering capabilities so our tools can help us run the business more efficiently.”
“Now we’re able to scale back on engineering and bring in new people with data science skills, giving them better data to make important decisions.”
The result: Oneflare continues to grow, boosting revenue
“We’ve doubled revenue since 2017,” stated Billy Tucker, CEO, Oneflare, in a November 2018 interview with the AFR. “We’ve now optimised the journey for multi-brand success and we acquire businesses because we think it’s more appropriate for us to have the right brand and business model for the right customer journey.”
“The support and guidance from Snowflake Analytics has given us confidence,” adds Nathan. “There’s a lot of trust: Snowflake Analytics lets us do it our way and helps us along the right path.”
With Snowplow Insights embedded as the core of its analytics customer data pipeline, Oneflare now has a clear view of business performance.
Tracking the entire user journey through Snowplow events gives Oneflare a much more detailed understanding of its customers explains Nathan: