Charlotte Langley, CMO at Bloom & Wild, breaks down how a handful of customer complaints turned into a permanent opt-out feature during the company's biggest sales week. She also explains how a three-person insights team uses AI to catch operational problems no human review would find, and why unit economics decide how far Bloom & Wild can push beyond flowers into new gift categories.
Behind the Expert
Charlotte Langley grew up, as she puts it, at L’Oréal — joining the UK graduate training programme and spending nearly seven years there, starting in sales with retailers like Superdrug, Tesco and Sainsbury’s before moving into marketing and landing on Maybelline. She left the machine for a smaller Australian makeup brand, leading its marketing in Europe, and then took a LinkedIn message from her now-boss that brought her to Bloom & Wild roughly six years ago. Today she is CMO of Europe’s leading D2C flower and gifting company — a family of three brands (Bloom & Wild, bloomon and Bergamotte) operating across Europe — where her remit spans brand, performance, customer insight and product.
The Quick Hits
- A complaint from dozens of customers, not thousands, was signal enough. Bloom & Wild built a permanent opt-out from Mother’s Day marketing — during the biggest sales moment of its year — and the response was enormous.
- Data-informed beats data-driven. Waiting for statistical significance would have buried the single most brand-defining product decision the company has made.
- You can’t replicate your first innovation forever. Decomposing letterbox flowers into what customers actually valued produced “care from afar” — a mission broad enough to expand against.
- Food and drink outperformed candles, and it surprised the CMO. Customers read flowers as a consumable that marks an occasion, so brownies and hampers were a shorter leap than home fragrance.
- Unit economics set the pace of expansion. You don’t need day-one margins, but you do need a credible path to them — pulling an unviable product back off the shelf costs you the trust that let you expand in the first place.
- A three-person insights team with AI can comb data no human could. It surfaced bursting hydration gel packs and proved a one-day-late delivery drives more dissatisfaction than any quality defect.
The complaint that wasn’t big enough to act on
Mother’s Day is the biggest moment of Bloom & Wild’s year, and the company markets it hard. Through its customer delight team, a small number of people said they didn’t want to be exposed to that marketing at all. Not thousands — enough. Bloom & Wild built the opt-out anyway, along with different flows for the customers who chose it.
This is the difference between being data-driven and data-informed. A purely data-driven team waits for the signal to clear a threshold; a data-informed one treats a weak signal from a high-empathy category as a reason to go look. Langley’s framing for why it matters here is precise: Bloom & Wild isn’t shipping a kitchen utensil that’s mildly annoying if it fails to arrive. It’s standing in for a customer with someone they care about.
“Care from afar” — decomposing an innovation you can’t repeat
Bloom & Wild was built on a genuine innovation: letterbox flowers, which solved the delivery problem in the UK. But as Langley says, you can’t replicate that first innovation forever. So the team went back and asked what customers actually valued beyond the flowers themselves — and found more replicable parts. Easy ordering. Reliable delivery. Real customer service. And, critically, an unboxing experience that genuinely reads as a gift.
That decomposition produced the mission the company now expands against: care from afar. Helping people show up when they can’t be there in person. It’s a useful filter in both directions — it tells Bloom & Wild what to build, and it tells them what to ignore. They are not trying to compete with ordering a book from Amazon and wrapping it for your grandmother. They’re doing the stand-in job.
Langley is candid that she got the first expansion bet wrong. Candles and home fragrance seemed obvious — flowers are fragrant, they live in your home. They sell, but mostly as an add-on. The category that actually took off was food and drink: letterbox brownies, hampers. On reflection, the logic is clean. Flowers are a consumable that carries an “I was thinking of you” message and then goes away. Customers made the leap to fresh food far more easily than to something that sits on a shelf.
Why unit economics decide the pace
Langley calls unit economics “maybe not the sexiest” part of the job. Bryan pushes back — they’re exactly what makes brand expansion possible, because trust is the asset being spent. Customers give a brand permission to move into new categories; launching a product whose economics force you to pull it back out of market spends that permission down.
The gating factor on how fast Bloom & Wild moves beyond flowers, then, isn’t customer appetite — NPS on the broader gifting range is running stronger than on flowers. It’s whether each new category has a credible path to economics that let it stay on the shelf. Expansion speed is a margin question dressed up as a brand question.
A three-person insights team, and where AI earns its keep
Research at Bloom & Wild started as a single UX researcher inside the digital product team, with everyone else fending for themselves. Professionalizing it took a director of customer insights and several years of teaching the rest of the business what research is for — including the discipline to say no.
The team is still only three people. AI is what makes it mighty: agents that scrape large volumes of existing data and push for hidden themes in recipient NPS feedback rather than just the themes someone thought to look for. The output is the same shape as the opt-out story — small signals that matter in aggregate. Over one Mother’s Day it surfaced hydration gel packs bursting in transit, an operational problem no human could have found by combing the data. It also showed that a delivery arriving a day late drives more dissatisfaction than any other quality problem — which turns into concrete decisions about checkout messaging and carrier investment.
Langley is equally clear about the failure mode. A tool they rolled out for trading meetings, deep-diving cohort models, turned out not to be working properly. Her response wasn’t to conclude AI can’t do this — it was to ask what let it reach the wrong conclusion, and tighten that up.
That lands on the bottleneck we hear about constantly: data integrity. Bloom & Wild wants to democratize access to its data without ending up with ten interpretations of the same dataset — because a frontier model will always hand you a confident, defensible-looking summary, whether or not the underlying data supports it. The risk isn’t that the answer is slow. It’s that you repeat it to the organization.
Retention starts with the product, not the CRM
Asked to unpack the retention engine, Langley starts somewhere unfashionable. Email matters, the app matters, occasion reminders and the rewards programme matter — but all of it sits downstream of whether the thing you sold is good.
Gifting adds a feedback loop most categories don’t have: the recipient tells the gifter how it went. That makes physical product quality a retention lever, not just a satisfaction one — “Bloom & Wild flowers last for ages” is a sticky, repeatable thing people say, and it traces back to supply chain decisions. It also makes recovery worth overspending on. When Bloom & Wild’s carrier tracking shows a delivery is stuck, they proactively resend before anyone complains. It looks irrational on a per-order basis; it shows up in repeat rate.
Sound Bites
- “We didn’t hear it from thousands of people, but we heard it from enough people to be like, there’s probably something bigger going on here. And then when we acted on it, we found that to be true.”
- “We have this data side, but we also have this huge amount of empathy for our customers and what we’re doing for them. And it’s the combination of those things which is powerful.”
- “Once you try and break down what people value into smaller parts, you then start to see the things that you can draw on for future innovation and future growth.”
- “Nobody’s expecting day one something to function in the way that it will in three years’ time, but you do need to know how you think you’re going to get there. Because otherwise you sell a promise of a product or a service that actually isn’t viable.”
- “Is insights what you need here? Do you need a survey? Do you just need to speak to customers? Can we actually answer this question? Or are you just looking for validation — in which case this isn’t the place to come.”
- “The important thing is that you don’t then go, okay, well, we can’t use AI for this. You go, okay, what did we do that meant it came to the wrong conclusion? How can we tighten that up?”
- “Sometimes as marketers we get obsessed with segmentation and personalization of marketing messages. But if the thing that you’re selling them isn’t good, then no amount of that is going to get them to come back.”
The Chord take
Two ideas in this conversation are usually treated as opposites, and Langley runs them together. The first is empathy at a scale below statistical significance — acting on dozens of complaints during your biggest revenue week because the category demands it. The second is unit-economics discipline that governs how fast you’re allowed to expand. Most brands pick one and call it a philosophy: the empathetic ones ship things they can’t sustain, the disciplined ones never hear the weak signal at all. What connects them is trust as a finite balance. Customer permission is what let Bloom & Wild move from flowers into brownies and hampers, and permission is spent every time a product gets launched and yanked. That’s also the clearest argument for why the AI work here is interesting: it isn’t generating anything. It’s making a three-person team able to hear the quiet signals — bursting gel packs, a day-late delivery — at a volume no human could process. Used that way, AI doesn’t replace the empathy. It scales the listening, which is the input the discipline runs on.
Put it to work
- 1Find the complaint you’ve been dismissing for low volume. Ask whether the size of the signal or the intensity of the moment is the better guide — especially anywhere your product stands in for a relationship.
- 2Decompose your founding innovation into the parts customers actually value. The replicable pieces — not the original mechanic — are what you expand on.
- 3Require a credible path to unit economics before launching a category, not day-one margins. Then treat “can this stay on the shelf in three years?” as a brand question, because pulling it costs trust.
- 4Point AI at the feedback you already have before you point it at generation. Hidden-theme extraction over NPS and support data finds operational problems no human review would surface.
- 5Audit data integrity before democratizing access. A model will produce a confident summary of bad data, and the real damage is repeating it internally as fact.
Full transcriptShow ↓
Bryan Mahoney
Welcome back to another episode of the Brilliant Commerce Podcast, where I get a chance to sit down with some of the brightest minds behind iconic — or as I like to say, next-to-be-iconic — brands. Today's guest will be no exception to that rule. Joining me today is Charlotte Langley, CMO at Bloom & Wild. Charlotte leads brand and performance and insight and product and so much more, and that's one of the many reasons I'm really looking forward to this conversation. Charlotte, welcome to the pod.
Charlotte Langley
Thanks so much for having me.
Bryan Mahoney
I already gave away the fact that I'm in Montreal today, actually recording this from the house I grew up in, so I'm feeling all kinds of nostalgia. Where am I finding you today?
Charlotte Langley
You're finding me in London, at home, hiding from the tube strikes. So I'm hoping we'll have a nice zen time and I won't be interrupted by too many people knocking on the door.
Bryan Mahoney
I'll keep my fingers crossed for you, you do the same for me. So, tube strikes aside — how's your week going?
Charlotte Langley
Yeah, it's going well. I was reflecting on all the different things that happen in a week. Yesterday we were talking about Christmas range confirmations and volumes. Today I've been talking with my leadership team about a whole process evolution for how we run the work in the team. Tomorrow I'll be prepping strategy share slides for the whole business. So it's very varied, but that's how I like it.
Bryan Mahoney
So you're keeping busy is another way of saying that. I found your background in the research I did leading up to this really fascinating. Before we get into Bloom & Wild, take a minute or two — tell the audience your story and what brought you there.
Charlotte Langley
Yeah. So I've been at Bloom & Wild now for about six years, but before I got here I kind of grew up, I always say, at L'Oréal. I joined their graduate training programme in the UK and I was there for almost seven years, and I worked across sales and marketing actually. I started in sales, working with big retailers like Superdrug, Tesco, Sainsbury's, and that was a super good grounding in what marketing is really for — which is of course to build brands and products that people want, but ultimately to sell stuff. I think we can sometimes be a little bit precious about that, so starting with that sales training was really helpful. Then I moved across into marketing. I worked mainly on Maybelline, in fact, so I still have a little soft spot for Maybelline — I still use that mascara to this day. It was a really great training ground; you learn a lot, they give you a lot of responsibility very early in your career, which is brilliant.
Charlotte Langley
But I just got to a stage where I thought, actually, do I want my boss's boss's boss's job? And the truthful answer was no. It's such a big machine, which makes it an awesome place to learn, but equally you get to a point where you think, am I personally making a really big difference, or could they just ship somebody else into my role and it would be fine? So I decided to make the leap to something smaller. I stayed in beauty initially and went to an Aussie, more startup makeup brand, leading their marketing in Europe. And then Aron, my now boss, dropped me a message on LinkedIn and said, I'm looking for somebody, are you looking? It was just really good timing. And yeah, never looked back.
Bryan Mahoney
I think those are the best ways of finding those opportunities. Sometimes they find you — it's just being open to it and being ready. And I like the intention behind the decision to go somewhere a little bit earlier, somewhere smaller, where it's less of a machine and it's more like, actually, we're building a company, we're building something durable. When I remark on the amount of time you've spent in some of those roles, I think there's not just durability in brand but durability in career. It's difficult to extract those learnings from fantastic brands and companies if you're only there for six months or a year. How do you stay long enough to make a meaningful impression, but also have a meaningful impression made on you?
Charlotte Langley
Yeah, you learn so much building something, and you go through different phases of a company and growth. I think that's really valuable — to know how to respond to different kinds of situations. And the other thing is, there's sort of a narrative that you need to jump around and get loads of experience, but I've had loads of different experiences in this one role. My job has changed every six to twelve months based on responsibilities I've had, what's going on. So if you are in a high-paced learning environment, then embrace it.
Bryan Mahoney
That's wonderful advice. There are a lot of things that are written that we sometimes believe to be true — like, it's been three or four years, I need to switch. Well, actually, you just need to keep learning. Sometimes you can stay within an organization and be curious and create additional impact. So the beauty background is really interesting. And then the opportunity at Bloom & Wild — what is Bloom & Wild? Let's start there.
Charlotte Langley
Yeah, not everybody will necessarily know the brand. So Bloom & Wild — we are Europe's leading D2C flower and gifting company. We actually have a family of three brands: we have Bloom & Wild, we have bloomon, and we have Bergamotte, in different countries across Europe. I'll probably talk mainly today about Bloom & Wild because it's the English-speaking brand and the largest in our portfolio. Our heritage is very much in flowers, but we're expanding into other gifting and that's kind of the future of the company. So that's a big part of the journey I'm now navigating with my team.
Bryan Mahoney
Going back through my notes, you mentioned something in our prep call that I wrote down right away — that a lot of the business is built on an obsession with customer and feedback. Going back to your days as a marketer and the feedback you got from understanding sales, as we think about putting that into operating practice in both the physical and the digital product, what does that look like today at Bloom & Wild?
Charlotte Langley
Yeah, it's super important. I'm always saying to the team, do we have any insight on this, and what's that based on? On the physical product side of things, we have some really good feedback loops through our customer delight team. So if there are problems, they come in from customers, and we use tech, AI, machine learning to categorize those so that we get really actionable outputs. We also have our NPS surveys — we have sender NPS and we have recipient NPS. Obviously that gives us very different data; when you're in a gifting business you've got two sides of that transaction which are equally important.
Charlotte Langley
All of that can help us pinpoint areas of our product range specifically that we need to change. We've just done our Mother's Day NPD for next year, and there's a particular rose type that caused a whole bunch of problems this Mother's Day, so we designed much less of that into the range. If we didn't have the data, we wouldn't be able to do that. Customers are telling us this is the problem, and we then have to make it into a readable format we can really action. That's the physical side of things.
Charlotte Langley
And then on digital product — the best example here, and forgive me if people have heard me talk about it before, is that we started allowing people to opt out of sensitive occasions. We're a gifting business, we're flowers, Mother's Day is the biggest moment of our year. We send a lot of marketing about it. We heard from customers, again through our customer delight function, that some people just really didn't want to be exposed to it. And what's interesting about this example is this is actually taking something that didn't impact very many people. We didn't hear it from thousands of people, but we heard it from enough people to be like, there's probably something bigger going on here. And then when we acted on it, we found that to be true. The response was enormous when we said to people, okay, we're going to let you opt out and we're going to build different flows for you. So there's a common thread here, which is really trying to aggregate lots of individual customer experiences into something bigger that we can then action.
Bryan Mahoney
I really like that. I think sometimes about the difference between being data-driven and being data-informed. If you were entirely data-driven and your hands were off the keyboard, maybe there wouldn't have been enough signal to allow people to opt out — well, we just have to wait for the data to tell us. But being data-informed and still trusting your gut and being connected to the customer, there was something that felt off, and it sounds like the team dug a little bit further. There was a really big impact for the people you listened to. I think that's where you get customers for life. That's what it means to be building a brand.
Charlotte Langley
Yeah, and it's so important in what we do specifically. You've hit the nail on the head, because I often say the beauty of Bloom & Wild is we have this data side, but we also have this huge amount of empathy for our customers and what we're doing for them. And it's the combination of those things which is powerful. We get that we're not just delivering you a new kitchen utensil, which if it doesn't arrive is a bit annoying, but you replace it, whatever. This is — we're delivering something which is standing in for you with somebody who you care about. And so therefore we have to think differently. We have to act on small things. We have to treat people as the individuals that they are and respond to their specific circumstances. Because if we don't do that, we're not providing the service that we promised.
Bryan Mahoney
It's like a proxy for you and your feelings about the person receiving it. Well, let's talk about that service for a sec. A lot of brands get started where there's a founder who can't find a product that specifically meets their needs, which is sort of a fancy way of saying innovation. And Bloom & Wild really started with a very real innovation in the UK — letterbox flowers solved the delivery problem. Another thing you said in the prep call that stuck out to me is that you can't replicate that first innovation forever. So as you're talking to your team, how do you think about continuing to expand categories? Your customers have given you permission to market to them on several occasions, they're looking to you for these moments, but you can't replicate that first innovation forever.
Charlotte Langley
Yeah, let me situate this in some examples. As I mentioned at the start, we're on this journey to expanding into broader gifting. And as we were thinking about how we do that, we had to really go back and understand, beyond just the flowers, what is it that people really value about our service and our product? And then you get to some much more replicable elements. So yes, it's how easy it is to order. Yes, it's that we have great customer service, that our delivery is pretty reliable. But actually you get things like the fact that the unboxing experience is really great for the recipient and it actually feels like a gift.
Charlotte Langley
That then becomes really key, because it drives two things for me and my team. One, what we're solving for here — we shouldn't distract ourselves with people who are ordering a book from Amazon and then wrapping it up and taking it to their granny. That's not what we're trying to replace. We're doing that stand-in job. So we call it care from afar. We're helping people show up when they can't be there in person, so we can very clearly focus on that mission and not distract ourselves with anything else. And then because we're focused on that, you see, okay, this thing has to really show up as a gift. And actually when you look in the market, it's quite hard to find people who really do that well. So you start to understand, this is what we can really build on — building that experience where everything in that box is intentional, has a place, it's a great experience for the person to unwrap it, take things out, experience it. Once you try and break down what people value into smaller parts, you then start to see the things that you can draw on for future innovation and future growth.
Bryan Mahoney
I love that. It's easy to assume that many aspects of the purchase journey, the gifting journey, the unboxing journey should be table stakes, and I think those are largely on the ordering side of things. It should be easy to find a product, easy enough to add it to your cart and move through checkout. But the act of giving a gift and having that gift be received — care from afar, I think, is a really great lens to look through — isn't necessarily table stakes. So how do you shift your customer's perception from we just send flowers, to we help people show care in a meaningful and reliable way from what could be hundreds or thousands of miles away?
Charlotte Langley
Yeah, that's the big question. We are very early in that shift, but I think it goes back to the idea of permission. I think we were almost too humble as a brand around what we thought customers would give us permission to do. And as soon as we started doing other things — the NPS on our broader gifting range is actually even stronger than flowers. People are really happy with what we're offering and we get great feedback. The thing we're very early in the journey about is re-educating people that we don't just do flowers. And I'll be honest, we're not very far in that journey. That's a big job that me and my team have to do over the next few years. We can talk later about how you invest responsibly and think about unit economics and all of those good things that are super important but maybe not the sexiest.
Charlotte Langley
But I think you have to have the baseline trust from your customer — that the first thing you said you were going to deliver on, you deliver on that, before you can expand into other things. I also think we've worked hard on the brand. We created back in 2020 the Care Wildly brand platform for Bloom & Wild, and that is really expandable to all kinds of different product types. People really understand what it means. At Mother's Day, for example, we've just done some out-of-home advertising on the tube, and my friends keep saying to me, you've gone viral on LinkedIn, because we shared this long-form ad copy that just speaks to what it's like to go home to your mum. The joy and the comfort of that and all the annoying stuff that comes with it, like fixing her phone, and her having moved your stuff out of your room that you thought was still yours — I'm saying this to someone who's in their childhood room. So through all of that brand behaviour and your comms, you build this sense of, this is a brand who gets it. And therefore I'm more likely to trust them to do other new things for me that I hadn't considered before.
Bryan Mahoney
This is a brand that gets me and gets it. And that's an incredibly difficult thing to measure, I would assume. It comes back to this idea of being data-informed and not entirely data-driven. And you said unit economics aren't sexy — I think unit economics are sexy. I think this is all commingled in this notion of trust. As your customers trust you and give you permission to expand into additional categories, you move beyond that initial product innovation. If you're introducing a product whose unit economics aren't strong, and you need to then very quickly remove that product from market, you're going to erode that customer trust. So to me it's really a formula that goes back to being a really smart operator, marketer, seller. You ultimately want to put product out into the world that customers love to buy and love to gift and love to share, and the way those products can be evergreen is you need to have strong unit economics. It's really about being a smart, responsible operator.
Charlotte Langley
Yeah, and you can have a path to that. Nobody's expecting day one something to function in the way that it will in three years' time, but you do need to know how you think you're going to get there. Because otherwise, you're right, you sell a promise of a product or a service that actually isn't viable.
Bryan Mahoney
You've done some testing into different categories, and we talked about this too. Talk to me about a category that surprised you the most — maybe where you thought initially customers wouldn't give you permission to go, and they were like, come on in, this is great.
Charlotte Langley
Yeah, so I'll be honest, I was wrong about this. When we were first expanding, I thought, okay, flowers, you put them in your home, they're fragrant — clearly candles, home fragrance, this is going to be what people want from us. And we do sell candles and people like them, especially as things to add on to their flowers, but standalone less so. We're doing better now when we put them in gift sets which have a really considered ritual around them. But actually the thing that is biggest for us is food and drink. So fresh baked things like letterbox brownies, or hampers.
Charlotte Langley
On reflection, the reason for that is probably that flowers are actually consumable. They're fresh, they're consumable, they don't hang around in people's homes, but they have this powerful kind of, I was thinking of you, impact. And so actually people transferred that to, okay, I don't necessarily want to send flowers for this thing, but equally I don't want to send them a thing that sits in their home. I want to send them something to mark the occasion and then it's gone. So I was definitely surprised by that. I thought people would find the leap from flowers to food bigger than they have done. Actually I think in a customer's mind it's like, no, it's a fresh consumable gift that marks an occasion well.
Bryan Mahoney
Yeah — I'm sending care from afar. Now when you say it, it feels obvious that they would look to you for that. Maybe initially, instinctively, I would say, well, that's not the normal place for us to expand into, which is interesting. So you have this feedback mechanism. The other thing I learned about you and the team that I thought was interesting is that you do have a small research team, and it seems like you use that research team as part of a decision engine. Did the instinct to build a research team come from your time at L'Oréal? Or was it just something that when you got there, you knew you needed — if we're going to do this product innovation, we need people doing this type of research and bringing us the ideas we can ultimately execute on.
Charlotte Langley
Initially research didn't sit in my remit. It's one of the many things I've picked up along the way, partly because it didn't really exist as a function. We had a single UX researcher, who you might imagine in an e-comm, tech-heavy company sat in the product team — the digital product team is where the research was sitting. And the rest of us were kind of fending for ourselves. So we really did need to professionalize that and bring in somebody who could create a function and provide some structure. Our director of customer insights has been with us for a few years now and has been doing a great job, but it's definitely been a journey for him. He's had to work hard to educate the rest of the business. Product were way ahead, they know how to use research, but only a certain type of research, whereas he's had to come in and say, hey look, here's all the things that we could do. What's on your mind? Tell me how you're thinking we might be able to help.
Charlotte Langley
But also really help them understand — is insights what you need here? Do you need a survey? Do you just need to speak to customers? Can we actually answer this question? What are you going to do with this insight, so is it actually worth doing or not? Or are you just looking for validation, in which case this isn't the place to come? So it's been a real journey over the past few years of getting more and more departments comfortable with what to use research for, how to brief them, how to think about actioning insights. But now it's in such a different place and it's really transformative, because it means we can really back up our decisions. And it's not just practice to be able to say to the board, we're doing this because of this — you also have much higher confidence in what you're doing and that you're not wasting people's time and resources. It's a small but mighty team. There's only three of them. And we're making heavy use of AI there to help them do more, extract more insight from the different types of research and data we have existing in the business.
Bryan Mahoney
I love that. The first part I love is just the discipline — just because you have a team doing research and producing insights doesn't mean that every single initiative needs research or can produce insights. Being intellectually honest enough to have the conversation, actually this one, we're not going to be able to help. Whereas a lot of times you have the department and the department is there to work — give me the mandate and I'm ultimately going to do it. I think that's really helpful framing. And then I'm curious about the use of AI, because one of the things I have seen is that as a thought partner, as a research partner, it can be really powerful if you have really detail-oriented humans processing that research to make sure we're actually getting meaningful insights out of it. Is there anything that's come out of the team's adoption of AI you might be able to share — maybe something that has worked well, or perhaps has worked less well?
Charlotte Langley
Yeah. One of the things I've been talking about with our insights director for a while is, how do we connect data and insight? We have a lot of data in the business that is, I suppose, under-potentialized, because it's difficult to access, difficult to read. So one of the things we've been working on is how can we use AI agents to scrape vast quantities of data. I mentioned before our recipient NPS data, for example. We were already using an AI-enabled programme to theme that feedback, but now we can do it ourselves and push it to look for hidden themes.
Charlotte Langley
It's great, because you get things similar to what I was talking about before — you get insights which are maybe not from thousands of people, but tens to hundreds of people saying things. This is very small, but a good example: over Mother's Day, at high-volume times things operationally always break down or go wrong, and one of the things was the hydration gel bags around people's flowers that were bursting or were not there. Now, we might not have picked up on that being a specific operational problem that we needed to fix with our warehouses, because a human would just never have been able to comb through that data. Whereas this way we can pick up on all of those smaller things that in aggregate have a really big impact. We've also been able to see, for example, that the delivery being a day late is one of the biggest drivers of dissatisfaction over any other quality problem. That's an interesting one for us to think about — okay, how do we message that better at checkout? What should we invest in delivery carriers, et cetera. So you start to get a different lens on things.
Bryan Mahoney
The ability to find those nuances in larger data sets like that — you're able to give this smaller team a superpower, whereas maybe they would only have been able to operate at a higher surface level before, and now they can get in there a little bit deeper. I still hear that wrangling all of that data into one place is still a challenge. We have so much data, but can we connect it? Can we get the right signal from it? Even with AI or machine learning or natural language processing — I think they're all close cousins — it's still difficult. We still need to have the data somewhere that's accessible.
Charlotte Langley
Yeah, but I think traditionally you just get bottlenecks, right, because you have experts who know how to interact with the data. And what we're working through as a company now is, okay, we need data integrity. So we do need our experts working on that, so that not everyone has ten different interpretations of the same data set. But there's a real democratization opportunity here, because if we can make those data sets more interactable with, if that makes sense, then you can really unlock a lot. So that's what we're trying to do at the moment — think about how does the data need to be surfaced so that the team can access this in the right way, don't make conclusions that actually don't stand up with a bit of scrutiny, but equally that we can remove those bottlenecks which just limit the amount of progress you can make.
Bryan Mahoney
This is literally the thing that keeps me up at night, and I haven't figured out the right way to tell the story, but I love the word integrity. Getting our data to a place where we have this high degree of integrity. My head of product has recently coined the term ARR — we sell a platform, and when I talk to the board the thing they want to hear the most about is, where's our ARR? But we're thinking about ARR differently, not in the form of revenue. When we think about the data platform that we have, we are using ARR to describe accuracy, reliability, and relevancy. And if we're able to deliver that as a platform, that is maybe one way of delivering data integrity. If you then deploy agents on top of something that has that integrity, I think you can start to imagine a place where I would feel safe democratizing access to that data and to those agents, because I would believe the output is going to be reliable.
Bryan Mahoney
The other side of the equation is that when we think about agents and these large frontier models, it is their job to give you some sort of answer and an opinion. And if the data isn't of high integrity, they're still going to summarize it for you. Then the danger is you're going to parrot that summary to the organization. It looks defensible, but it might not be accurate, and that can have real ramifications. So I still feel like we're in the very early innings. These tools feel so magic — you ask for something, you get an answer in a time horizon that was previously unimaginable. But remembering to slow down a sec, let's make sure that we have high integrity, that the answers we're getting are the right answers, is a reflex that requires some amount of maturity and experience. There's no way around it.
Charlotte Langley
Yeah. And trial and error. We rolled out something that was helping us with our trading meetings, deep diving into cohort models and stuff, and then we uncovered that actually it wasn't quite working properly. But I think the important thing is that you don't then go, okay, well, we can't use AI for this. You go, okay, what did we do that meant that it came to the wrong conclusion? How can we tighten that up? And also, things are moving so fast. These models are moving so fast that stuff that was, as you say, unimaginable even three months ago is now possible. So you've got to keep testing, keep iterating. It's a bit of an exhausting time because it's so fast and you've got to evolve really fast. But I firmly believe that we will get to a point where we trust the output, and it's just about working out what we as humans need to put in place as the parameters.
Bryan Mahoney
Exactly — so that we can get to this place that is the holy grail, where we can legitimately democratize access to data to the entire organization. Wouldn't you love to have people be able to have a hypothesis, have an idea, have a question, and get a really good answer? Because then that starts the whole idea flywheel within the organization. But when there are those bottlenecks — historically that's just been a difficult thing. Years ago it was, if we invest in business intelligence then anyone can go and self-serve, but that really hasn't played out. I do think we're on the cusp of AI being able to deliver on that promise. I don't think we're going to get there as fast as most people feel; there's still an awful lot of hard work to be done to get to this place where we can trust, and where there is that integrity. But it would seem to me that you have a leg up in the fact that you've had research and insights as part of the foundation of what the team does, and this idea of, okay, let's test and learn, let's be really curious, but let's be really thoughtful and careful. Because going back to where we started, your customers have given you permission to test along with them, to bring them along in the journey.
Charlotte Langley
Yeah, exactly that. And we keep going back to them as we go.
Bryan Mahoney
Well, let's use that as a segue to talk about retention. I'll say something that is super obvious — chasing demand is really expensive. So instead of chasing demand, how do we think about creating demand? And creating demand with customers we already have a relationship with is a really wonderful place to start. I'd be really curious if you wouldn't mind unpacking the retention engine at Bloom & Wild. I think product quality is a part of it, the quality of the relationship you have with your customers. What are some of the things that make the engine go?
Charlotte Langley
There are so many parts to it. Sometimes people think retention engine, CRM — and that is a big part of it, and don't get me wrong, email is a big channel for us, but there are so many other things to it. We have to deliver a great experience in the first place. There's the digital experience we've talked about, the physical product. And what we're very acutely aware of is that what people say about the physical product we delivered is really key to retention, because there's a feedback loop between the gifter and the recipient of the gift. If you send me flowers, I'm going to send you a WhatsApp and say, hey Bryan, thanks so much for the flowers, they're beautiful. Or if it doesn't go well, I'll say, thank you so much, but you should know that actually they didn't arrive well, so you should get them to give you a refund.
Charlotte Langley
So for us there's another feedback loop, between the recipient and the gifter, and that's really important in terms of what's memorable for people. One of the things we find — this is kind of anecdotal, but it also comes out in the data — whenever I talk to people about Bloom & Wild, they say, I love Bloom & Wild, the flowers last for ages. So that's a memorable, sticky thing about our physical product that is different from another brand they could get it from, and that is because of the way we've set up our supply chain, et cetera, which I won't bore you with. You never forget in marketing that you're selling people a product or a service, and that thing has to be good in order to drive retention. That sounds so basic that I almost hesitate to say it, but I think sometimes as marketers we get obsessed with segmentation and personalization of marketing messages. It's like, yeah, but if the thing that you're selling them isn't good, then no amount of that is going to get them to come back. So that's the first thing I would say — the proposition. Is the product and service right?
Charlotte Langley
Part of that for us is our customer delight team generosity. One of the things that drives retention for us is that if things go wrong, we fix it really well. I'll go back to Mother's Day because it's fresh in my mind. If we track deliveries through our carriers and we can see that something's going to be late, we just proactively resend it. Nobody's asked us to do it, but we just know it's important and we're like, this needs to arrive, so we're going to send another one because that delivery looks like it's stuck. You might think, well, that's crazy from an economics point of view, but actually we can see that then translates through to repeat rates, because people are so impressed with how we've dealt with a problem. That's a really important part of the flywheel.
Charlotte Langley
And we've talked about brand as well — tone of voice, feeling like a brand that gets them, that drives retention, because people feel like they're part of something, that they're contributing to a business they share some values with. And then you get to all of the more functional stuff, like your triggers, your nudges, your promos. We have occasion reminder functionality — we encourage people to save occasions, we remind them that someone's birthday is coming up, give them an offer. We have our app which makes it really easy to repeat, saves your addresses, all of those things. We now have a rewards programme as well, which is proving to be really meaningful in terms of driving repeat revenue. Those are all the marketing mechanics, but I mentioned a ton of stuff upfront which I think sometimes gets forgotten.
Bryan Mahoney
It gets forgotten because that's the hard part. Actually, I don't think it gets forgotten — I think it's easy to start with the pieces that feel like a little bit of blocking and tackling. And I do think that as an industry we're sometimes like, well, AI is going to go and solve all of these things for us, the nudges and the reminders, and I can reinvent my CRM. But no matter how many of these conversations I have, when we look at what it means to be a great and durable brand, it comes back to: is the product fantastic? Do you obsess over making your customers love you? Those are the really hard things. What I've loved about this conversation is that you've given us real insight into how you're using technology, even how you're leaning into AI, to make some of those blocking and tackling functions a little bit easier and more effective. But at the end of the day, if the product isn't great, and if the experience your customers have with you as a brand isn't fantastic, it's not going to matter. No amount of AI or technology is going to help us out there. I want to keep repeating that message, because that's the hard part, and to me that's the age-old equation for building something durable and lasting. That's exactly the message you delivered to us today.
Bryan Mahoney
So maybe that's a good place to wrap up. This has been really fantastic. I've loved getting to know you and getting to know the brand, and I'll be continuing to watch and see what additional categories you all are expanding into. Charlotte, thanks so much for joining me today.
Charlotte Langley
Thank you, it's been a pleasure.