Balsam Brands CTO France Roy sits down with Multichannel Marketer to discuss how BalsamHill.com will better use its data to identify high-value customer segments and drive higher conversion rates.
Balsam Brands chief technology officer France Roy needs its data and analytics team to do more than just pull reports.
The team needs to easily access its metrics and help its business teams make data-driven decisions, like identifying high-value customer segments and then market to them appropriately.
Balsam Brands chief technology officer France Roy
But when its team has to use a SQL database to extract data, it could take months to pull data for its marketing, merchandising or operations teams.
The artificial Christmas tree owner of BalsamHill.com decided to invest in its ecommerce data tools in a quest to offer better personalization and drive higher conversion rates.
The retailer operates mostly online plus a handful of physical stores. Nearly all of its revenue is related to the Christmas season, and mainly generated from artificial trees. But it does sell other items, such as ornaments and wreaths, and other seasonal items, such as Halloween decorations.
With such a heavy focus on the holiday season, the brand spends the first three quarters of the year implementing and testing new technology — including data platform SoundCommerce and data cloud Snowflake — before letting it loose for the all-important Q4.
The first proving ground is its annual Christmas All July sale. While the 2024 sale did not surpass last year’s event, the merchant was still pleased with the results, Roy said.
Roy recently sat down with Multichannel Marketer to discuss the brand’s new technology initiatives, previous pain points and how this will drive increased conversion during the 2024 holiday season.
Multichannel Marketer: How did the Christmas All July sale go?
Roy: Overall, the market for home goods, our whole category, was down. Luckily, we weren’t down as far as the market overall, which I think is due in part to a lot of efforts we’ve had over the past six months around uplifting our user experience, bringing in some personalization tools.
We implemented a personalization algorithm that basically detects the user’s intent on the site. We have a lot of different users shopping for different things. Some are coming for home decor, decorations for the season. Some are coming to us for having great deals on Christmas trees. The better we can figure out what is the customer there for and how do we craft an experience around that on the homepage and the product listing pages, the higher conversion we’ll see.
We’re starting to thankfully see some of our efforts to increase conversion across the site come to fruition.
MM: How does the tool measure intent?
Roy: It’s based on the data that we have of the customer of which we’re collecting from various sources. So if you come from a product listing advertisement that has, for example, Halloween decor on it, then we know that you are interested in the site because of that and thus you’re maybe less interested in buying a tree.
For us, trees are the lion’s share of our revenue. In the past, it was trees upfront. If you were a customer of ours last year and subscribed to our email list, every single email you got would’ve been about trees because that’s what we want to drive. But in the world of one-to-one marketing personalization, and now this next generation of AI tools that’s just coming available, we’re looking at how do we leverage that.
MM: Can you give me an example?
Roy: For something like, ‘What was your ad?’ Maybe we were targeting a certain demographic with those ads. We will now have that data and then we can see based on browsing patterns on the site, what people look at, what they search for. And then what they actually converted and bought at the end.
So, this gives us some threads we can then tie through as we get to the core season to say, ‘Here’s some learnings we have that will help us get bigger conversions.’
We did have an over 20% lift in Christmas All July in our non-tree decor. Normally, we really push trees. So now we can see, ‘Oh, actually they’re interested in other things. How do we feature those on a site in a way that drives more conversion for them?’
MM: Someone who buys a tree and then you’re going to market to them with all the trees, but they just bought a tree? So do they need another tree? How do you guys handle that with your business being so focused on trees?
Yeah, it’s been interesting. There’s so many pieces to it. I personally bought my first Balsam Hill tree during Christmas All July. And I found it interesting that in the shipping email that we send out, there’s a personalized, “You may also be interested in these other items” at the bottom of that email. And one of them was a branch sample kit, which is usually something you get before you purchased the tree. And I had already bought one from Balam Hill. That was just one example where we’ve got a misfiring in the system.
What it’s going to take is going through each piece of our customer journey and really discovering what are the customers really doing here on this trip, and are we giving them what they’re looking for, and how can we be constantly doing that better as we move along.
MM: I’ve heard you upgraded to SoundCommerce and Snowflake. Can you talk about how this data you are getting from these new tools feeds into these platforms?
Roy: The platform, we have tons of our systems plugged into it. Before we were basically operating on a SQL database. There was no real global data warehouse like Snowflake. We are using SoundCommerce for their models and all the capabilities that our merchandising team, for example, can actually go in and find usable data that’s structured in a way that’s easy to access and can bring fresh insights to the team.
MM: What is a marketing example of how you hope to use this?
Roy: One of our key initiatives is going to be how do we get returning users because we have a lot of people who come to buy a tree, but that’s all they would buy. And so can we follow up with ornaments or storage sets or different things. And, how do we craft those customer journeys in a way and then test different journeys against other ones to see which drive the highest conversion. That’s the direction that these tools are allowing us to move in, and we’re doing it very iteratively as we go, trying to take one piece at a time, learn and then build on top of that.
MM: What’s been one of the largest benefits?
Roy: I would say the coolest thing is the ability for us to simply integrate with other systems as we see fit. We were just looking at some competitor analysis tools that show us competitor sales, similar category sales on their own ecommerce sites, on Amazon and so forth. And we can grab those from their API and easily integrated into Snowflake, connected to the SoundCommerce models. And that just gives us a capability set that in the past would have been very difficult to bring in-house and even more difficult to bring that up to the users in a way that they can see it and access it easily. And now it’s just so much more streamlined with all those systems.
Now, anytime we have some new data source, let’s say we’re bringing on a new marketing attribution provider. We can then look at those attribution metrics and decide which way do we want to do it, how do we pull in the data so everyone can access it, how can we tie it to other data? All of this was very manual in the past and now we’ll be able to automate it in a very streamlined way.
MM: How is this different from what you were doing before?
Roy: The team has an idea of what they want. They’ll ask this other team to go build it for them and come back. And the back and forth to just get what they really want takes so long.
And now we’re trying to position it so the business users have data at their fingertips and can make those decisions in real time versus waiting weeks or months to get the right data to then make a decision.
When we have what we call our ‘million-dollar days’ during season, we don’t have a lot of time to make adjustments. Every single day counts. And the more feedback we can get in real time, the more leverage we have to drive revenue conversion and so on in the business.
MM: It really would take weeks or months to get the data set that you would need?
Roy: Depending on how complex it is, yeah. Because we didn’t have a data warehouse. Like I said, we had a SQL database, so if you want to link up some data, you’re literally going into 10 different tables and trying to do complex SQL joins. And then where SQL doesn’t work with the team was writing Python scripts to manipulate the data around to then create the report.
With SoundCommerce, just having data models defined and they’re defined in a way that aligned to our category, they’re retail data models. It allows us to just much more quickly and easily get the data we want, and have it all in the right format. That’s a process we’ve done in implementing SoundCommerce is cleansing all of our data.
For example, we switched our order management systems a couple of times in the past years. So the format for the order data was different depending on the year that it came in and what changes were made during that year. So trying to bring that data to then write a report was very painful; versus in SoundCommerce, it’s just one model. Everything is ingested, cleaned up, and then any new data set that needs to tie that model, we have one place to go and do that from.
MM: And this is with your data analytics team?
Roy: Exactly. And we’re trying to not just give our business users more data, but also integrate those data analytics teams more deeply into those teams so that we have a data analytics person, for example, thinking about marketing or thinking about ecommerce conversion, not just producing a report, which with our old infrastructure, was very challenging to do that. So that’s where their focus was. And now we can really focus on adding value to the business users and not just creating the report itself.
This interview has been edited for length and clarity.
Network
