Dennis Hegstad is cofounder of LiveRecover, the ecommerce SMS business where real people (Dennis’s team) communicate with potential customers via SMS to win back abandoned carts. He’s grown LiveRecover to run across 2,000 active ecommerce sites over the last 2 years. Prior to that he ran a viral blog with more than 30 million unique visitors, ran his own network of Twitter accounts with 9 million followers, and ran Fashion Nova’s paid Instagram & Facebook native ad activity.
In this interview he talks about his experience across this, how and why he launched LiveRecover, and what they’re doing on their journey toward $10 million annual recurring revenue and beyond.
Dan: Hi, Dennis, thanks so much for chatting today. Hopefully this will be interesting and useful for lots of people working around ecommerce. First off – what’s your background and how did you get to where you are just now?
Dennis: I started really early on in ecommerce. I mean, even before ecommerce I was playing Counter Strike and other games – not professionally, cause it wasn’t really a thing back then – but I was sponsored and played that full time back in 2003 and then from there I got into MySpace and then I started a t-shirt brand/apparel brand on MySpace in 2008.
Fast forward I moved to Los Angeles and then I got into building web products. So I did that for a couple of years – I worked at a startup called Mahalo. The founder was a guy named Jason Calacanis – he’s @jason on Twitter.
Dan: Ah yeah, he’s well known.
Dennis: He’s a smart guy and he’s sold some companies for a hundred million plus I think, and made some really good investments, but I worked there a while, and while I was there I was also building up a bunch of Twitter accounts.
So I had like maybe 9 million followers on Twitter after a couple of years, which were all different accounts that were themed like cars and girls and food and travel pictures and houses and architecture and memes and whatever. So I used that to build up a lot of web traffic. I mean, I could drive50,000 uniques to a hundred thousand uniques a day.
So from that I built up a website that was like a viral blog. Within the first 90 days we did 30 million uniques and 250 million page impressions. We were monetizing through AdSense. From that then we built up a marketplace, started onboarding influencers and built that into a business. I kind of left ecommerce behind for a bit.
Then, a few years later, I got back into ecommerce – I sold all my Twitter pages off, got back into ecommerce and doing paid media on Facebook.
Dan: Your own ecommerce sites?
I ended up working at Fashion Nova – I consulted at Fashion Nova – running their Facebook & Instagram native paid ads in 2017 for about eight months. Then I left because they wanted to hire a full-time employee.
Fashion Nova is one of the US’s largest fashion ecommerce sites, they grew 600% in 2017 – the period where Dennis managed their paid ads – according to their CEO
Dan: Then from Fashion Nova, that led you toward building LiveRecover?
Dennis: Yeah. Then a couple of months later – maybe six months later – I started building LiveRecover with my now co-founder. We’ve been doing that for about little over two years, two and a half years now. And we have, I think nine or ten thousand merchants have installed our app at this point. I mean, we’re not actively servicing 10,000 merchants, but that’s how many have come and gone. We have a couple of thousand now that are monthly paying customers, and they’re live with us.
So that’s it – the last 10 years, 15 years in a nutshell.
Dan: That’s super interesting. I guess it’s gone fast. Just jumping back before we talk about LiveRecover, there were a couple of points there where it sounds like pivotal things happened. The Twitter accounts – what did you do? Did you buy them and then build them up further with the idea of selling them on?
Dennis: No, I built them all up from scratch. I built them from zero. I just posted all day, I would schedule with Buffer. I would schedule out all my tweets on a Sunday. I would pretty much watch movies, get drunk, and schedule tweets on a Sunday. And I would spend five, six hours and schedule the next two weeks of content. And then every Sunday I would at least give myself some leeway. If I was lazy, then I could give up, get a week off. But if I wasn’t, then I’d always have a little bit of a backlog. And so all I would have to really do is log into those accounts, retweet my accounts across each other and kind of daisy chain engagement. And then the manual work at the beginning was putting the ads in right. Tweeting ads for other companies.
So yeah, I built them all up from zero and then I sold them to an app company later in 2018.
Dan: That’s interesting. Why did you want to sell them?
I just thought like, I’d rather build software than manage ad accounts. Twitter was kind of dying in terms of having the most engagement and all the attention. If you looked at the market, I guess in terms of Twitter versus Facebook, Facebook acquired Instagram, Instagram got so big, then WhatsApp. It was like Instagram kind of made sense. If you look at like what was happening on Instagram in comparison to Twitter, there are meme accounts on Instagram that built hundred million dollar a year media businesses like @fuckjerry and @thefatjewish. And there are all the alcohol products and direct to consumer products and brands that build and sell and distribute through Instagram – Instagram is the place to be now. So it was kind of looking for like a, not a way out, but I’d rather offload it and get some cash and reinvest that into a home or a car or a new business or something. So that’s what I did.
Dan: That makes sense. So basically you didn’t want to go across to Instagram and build up the whole thing all over again.
Dennis: I felt like I was getting old. I mean, I’m in my thirties now. So at the time I was still in my late twenties I think. 29 maybe. And so I was kind of, it kind of felt like closing the chapter for me.
I would have been smarter to hire someone just to run all the accounts for me and then just monetize them and probably make tens of thousands a month off of the network of pages. And then it would have been a ‘smarter’ decision. But again, I had a network on Twitter and not Instagram.
Maybe I should have just bought a bunch of Instagram accounts, but either way I was thinking: is this an interesting thing to do?
And on the money longer term: You can make a bigger multiple on selling a SaaS business or a, or a marketplace, or some kind of software product than you can by posting to an Instagram account that could get banned at any minute by their terms of service change or something
That’s why I like SaaS. You can make money while you’re sleeping. You don’t have to worry about anything like Instagram taking that away. Yeah.
Dan: So was that the motivation? You went into it thinking “okay, this is going to be a much better business model and I can build something up and maybe sell it in a few years”, or something else?
Dennis: With my business partner – I was showing him some of the stuff that we were doing at Fashion Nova – some stuff which was public facing. At Fashion Nova at the time we were doing things with Facebook messenger, retargeting and chatbot stuff. And then he came to me and was like “some of the stuff that you showed me gave me this idea”.
And I looked and there’s already validation in the marketplace, which was SMSBump and Attentive, which I think were the only two players at that point. But even then, Attentive had raised $10 or $15 million already. And they had just launched, and SMSBump had already been in the market for a year. And so he’s like, “Hey, we should do something like this, but take a note out of the affiliate marketing book, where a lot of lead gen sites will have live call centers, and call you to close leads”.
So we were like, let’s just do a similar thing, but over text with live humans. The goal was to obviously build a really good tool to help merchants on Shopify, and then eventually all across ecommerce platforms. But, obviously to make money too. I had never done a SaaS business – and my business partner, I don’t think had either at the time. So we were like, “Hey, this is going to be a challenge”, but it should be interesting because, long-term, if we get to – say – $10 million ARR or $25 million in annual recurring revenue, you don’t have to worry about if your ad accounts are getting disabled on Facebook or if your supply chain is messed up or there’s something wrong with your products, there’s no inventory or cost of goods. It’s like you build it and then you just continuously sell it.
There was a lot of appeal and to trying a SaaS business, but definitely took us some time.
We didn’t pay ourselves for the first eight months.
I’m sure not everybody is in a flexible enough position to be able to bootstrap. It wasn’t as fun and glamorous as it sounds. Now, it’s great because we’re doing well. But at the beginning we weren’t taking any money out of the business.
Dan: Yeah. That makes sense. How do you feel now about having bootstrapped things and succeeded?
Dennis: Well, I mean, I see some businesses that are doing good and they’re make a million dollars a year, but they’re spending a million dollars a year on their employees and all their salary and all their other cash burn. And it’s like, well, that’s not very interesting. I probably wouldn’t pay myself until like the company was profitable enough to pay me twice. Not until we have excess capital & we’re profitable.
Dan: You didn’t pay yourself for the first eight months. That feels like a lot of time to go without money coming in, but not a long time in terms of building software. So presumably you built everything pretty quickly. How, how did you go about building?
Dennis: My business partner was on vacation in, I think in Italy or Switzerland and he was like, “Hey, I’m going to start this kind of before I leave and finish it when I get back”. And then six weeks later, everything moved on and our basic MVP was ready to go.
I was connected enough – because I’ve been living in Los Angeles and working in ecommerce and paid media and doing the influencer marketing stuff. I had a handful of brands that were willing to try us in our beta, like within the first day of launch, we were ready to onboard stores. So we got some decent brands, maybe 15 in the first couple of days. And then from there, we were making a couple of thousand dollars a month, I’d say right out the gate, which isn’t much, but when you’re trying to grow, you can’t just pay yourself when you’re making that little.
So we reinvested into things like optimizing our sales copy, running Facebook ads, doubling down on, on different things and tools to make our lives easier. After eight months we were at a point where we’re like, cool, we can create a payroll and have real employees and do this.
Dan: You said you started with about 15 brands. How did, how does it work from a technical point of view? How do you onboard a brand?
But yeah, Shopify is the, is the main source of our stores. We’re just now starting to sort of expand horizontally into WooCommerce and Magento and Cratejoy and things like that.
Dan: That makes a lot of sense. So the first 15 were Shopify merchants that you knew, and so they could just plug you straight in.
Dennis: Yeah. Up until the first two years we were all Shopify. Now we have maybe 50 to a hundred WooCommerce stores. But the 10,000 that I mentioned we’ve had come and go were 95% Shopify, I guess even 98% Shopify.
Dan: Got it – so it was basically a Shopify App first from your customers’ point of view, but now you’re available for pretty much any ecommerce platform. I guess starting out on Shopify informed lots of decisions at the start. How did it work from a pricing point of view at the start, and how did you position the value?
Dennis: At the beginning, we didn’t really know what we were going to price our product at. So we just did performance-based pricing where it’s literally zero risk. You just come, we take a commission on what we recover for you.
And so it’s like, if you say we take 5% commission and we make you a hundred thousand dollars, well, you give us $5,000. If you’re a media buyer and you’re comparing against ads, that means you get a 20x return on ad spend, which is a great metric. If you’re getting that on retargeting, you’d probably be drinking champagne every morning.
I think for us, that was a great way to position it. And then we made it tiered. So 5%, if your average order value is this high; 10% if it’s this high; 20%, if it’s this low. Then we changed it where you could pay a monthly subscription fee to essentially change the tiered performance rate to the commission.
That meant it didn’t matter about your average order value. You could pay $0 to work with us. This is what we have now:
- It’s $0 a month. No monthly fee, but you give us 20% commission.
- Or $49 a month and we take 10% commission.
- Or $99 a month and we take 5% commission.
So that’s kinda what the pricing structure is now. In the past, where we were free, if we didn’t have that level where we could see whether someone was willing to pay $99 a month to get the 5% pricing, it was almost very hard to determine what someone was worth to us. So customer acquisition was pretty tough. But once we put a paid plan, a sort of monthly pricing plan in place, in addition to the commission, it made it a lot, lot easier to start scaling.
Dan: And talking about scaling, how do you find customers? How did you find customers at the start and how has that changed to now?
Dennis: I think now we’re kind of at a point where we’re starting to hire sales, business development reps and content marketing people, marketing automation people. So if someone is reading this: if that’s you, come hit me up and talk to us.
But at the beginning, it was all just me and my co-founder doing anything scrappy, right? And we’re still doing cold email outreach, of course, but I was messaging every one of my friends on Instagram and Twitter nonstop, talking on groups and jumping into conversations and letting people know about our tool almost maybe a little bit aggressively. But if you don’t want to spend money at the beginning and you want to validate your product, you gotta do what you gotta do. But we definitely relied a lot on Facebook and Instagram ads because starting out we were a private Shopify app, meaning we were not in the app store.
So if you searched for us in the Shopify app store, we weren’t there, we were unlisted. We’re still approved by Shopify, but, um, you would have to come directly to us. So we never got organic exposure sort of the equivalent of SEO value within the ecosystem of Shopify because they would never blog about us; they would never give us any backlinks, nothing.
Now we’ve submitted and we’re approved be in the app store publicly, but at the beginning it was all Facebook, Instagram ads, direct outreach, word of mouth. And I think our strategy is kind of the same, up until now where there’s a little bit more competition in the market now. And so now we’re proven and we’re bigger, we can afford to double down on some more resources like building out sales and business development teams, account managers, stuff like that. Cause right now we’re only a team of about, I think we’re 11 full-time or 10 full-time.
Dan: That’s interesting. I can think of a few SaaS businesses that have started out like that, almost opportunistically, with very low competition at the start and then over time that gradually crept in and they built up sales teams and become large enterprises.
Dennis: You look at Klayvio right. I mean, they’re an email abandoned cart abandonment tool, and they’re more than just email for cart recovery, but they’re focused on ecommerce. I mean, they’re a two and a half billion dollar valuation business. So crazy.
But there’s a lot of people getting into SMS too now, right? Like even Klaviyo are now turning on SMS, but, we’re a little different, that’s why we do the human powered thing versus the automation part.
Dan: That feels like a great sales pitch, because it makes a big difference having actual humans do something like SMS rather than it just being automated – it feels quite different to email.
Dennis: It does. It’s actually funny. When we have sales calls, I usually tell this story because, whenever the conversation comes up, people say “Hey, do people get ever get upset about getting text messages or do they like forget that they gave their phone number and get mad?” – And I’m like, yes, sometimes, but it’s pretty rare.
When it does happen, there are some scenarios where people will start replying telling the ‘chat bot’ to go away. And then our rep – our texting agent – will say “Hey, I’m sorry, but this isn’t a chat bot”. And then they’re like, “Oh my God, I’m so sorry!”. And then occasionally you’ll get someone who says something like “Hey, if you can prove that you’re real, I’ll buy from you”. And then our texting agent will send them like a meme of like a bunch of dollar bills, with a message “Give me all your money” or something. And then they’re like, “Oh my God, this is amazing!”. And so we’ve actually had situations where people who may have been put off because they thought it was a chat bot and that was the expectation. And when they found out that it wasn’t, they actually converted into a customer through the personal service.
Obviously that’s not going to happen every time. But I would say even like half the time when those few people are upset, when they get a real human response, they feel empathetic. Almost like, “Oh, I feel bad for the way I was just acting”. Not all the time. But I do think the human part matters a lot more, especially because texting is sensitive. I think copy and pasting your email strategy straight over to SMS is very lazy. It’s like occasionally we get criticism from people saying “Oh, duh, just copy and paste all your flows to SMS”, but people don’t want to be texted the way that you emailed them.
Dan: That makes perfect sense to me. Emails and texts are very different – both the way they’re written, and the kind of expectation. Different in different countries too, which is an interesting question: Do you think that the business is very USA-centric, or that you’ll push out into different regions? How are you planning that?
Dennis: So we we’ve stayed US and Canada focused only for now. So we can lean into relying on English-speaking texting agents, the, the rules in terms of compliance and how texting works. We understand the landscape here and in Canada, and then the cost per texts: we don’t charge per text message. The cost is just significantly higher in all those other countries. And so for us to operate there and make it profitable for us, we couldn’t keep the same pricing model. So there would be like price discrepancies, like it’s $99 a month, plus 5% in the US, but it would end up being something like $99 plus 7.5% in the UK. It’s like we’d have to increase the commission or something to make it as worthwhile. Right? Plus compliance and privacy plus language localization add to that.
There are other issues too: If we want to hire agents that speak Norwegian, and Swedish, that’s going to be a lot harder than finding someone who can speak German or someone who could speak Dutch or someone who could speak Danish, so covering Europe there’s just too many moving parts compared to North America. So I think for now, at least until we get to a certain level, we’re just going to focus on kind of North America, but we definitely want to expand. It may be that we’d probably expand through WhatsApp most likely more for the rest of the world; more so than SMS.
Dan: Do you do any WhatsApp stuff at the moment?
Dennis: Not yet. We definitely have been debating it for about a year. The WhatsApp API availability hasn’t been open enough until recently. And now it’s actually ready to start being used.
Dan: So WhatsApp is basically something to test in the future?
Dennis: Yeah, definitely. In 2021, we’ll probably be testing and putting out a channel at least for WhatsApp.
Dan: I guess that’ll be really interesting. Product-wise, do you have a roadmap, or is it less formal than that?
Dennis: We have, I mean, we do have a roadmap, but we don’t have goals or ambitions to build an all-encompassing SMS platform. If that makes sense. We want to build a great product to solve a particular problem.
Attentive, Postscript, SMSBump – or maybe I’d really just say Postscript – have built the MailChimp of SMS. And they’re a great “at scale” automated tool for lots of things. They’re like the Swiss army knife.
We don’t want to do everything – we just want to be the steak knife – the best at recovering abandoned checkouts, because we have live people who can close the leads for you with those abandoned carts in real time. And most people, most brands don’t have the resources to have someone in real time, so the part we do, they cannot do that with just software.
The example I usually see is you go to a website and they have an Intercom chat bubble on the bottom of their website or drift, or whatever live chat tool, “Hey, click here to chat with us”. Right. It’s like Friday, 2:30 PM. And you’re like “Okay, cool. It’s 2:30 on Friday. I’m gonna chat with the brand”. So then you click into the bubble, and it says, “Sorry, we’re out of the office until Monday – leave us a message”. And so you leave. Why is that chat bubble even there? That should be hidden if you’re out of office, there’s no reason to have something that looks like live chat when you don’t, unless you’re trying to collect leads, but even then that’s a terrible experience for the customer.
If you’re asking them to chat with you and then you don’t chat back it’s not going to help. So I think with SMS – that’s where we want to focus. Being human at the most effective time possible. That experience and the environment of, “Oh, I just abandoned something. My intent is really high. I got to the stage of putting in my information and then I left and then someone’s willing to offer me the personalized service to try and complete the sale”.
It’s like you’re in a shop and there’s a personal shopper who spots you walking out the door and is like “Hey, Hey, sorry. What can I do to get you to come back?”
That’s the equivalent of our agent writing a post. And I know that there are lots of tools for automated emails coming because people didn’t complete checkout on websites. But with live people sending an SMS, it’s just a totally different experience. So the roadmap is very much focused around us, enhancing the tools that we have, which is cart recovery through SMS.
We do have discussions on rolling out a churn recovery product. So if you’re a subscription business powered through Shopify or Cratejoy, or WooCommerce and someone cancels their subscription, we’d like to send them a text and ask them to come back for some kind of discount or learn why they churn. And then there’s a plan to have a live chat feature via live SMS. So if you haven’t added to cart, right, you haven’t made it to checkout and abandoned, but you’re on the product page, and you have a question about something the same way. Instead of that Intercom or Drift chat popup, instead of that you can text us. And that, that function would allow you to text in and have our agents respond and talk to you live to give you pre-sales support. So those would be the only products that we’re planning on really ever building at the moment.
That doesn’t mean that we won’t do more advanced analytics and other features to optimise things. Like we just added AB testing for coupons. So if you want to try $5 off versus a 15% offer $5 off versus free shipping coupon to see what gives you better results, we do that. And we are building more stuff like that, but we’re not going to build full automated SMS campaigns. We’re not going to do browse abandonment and follow-ups and automated drip campaigns and flows. That’ll never be a part of the business.
Dan: That makes good sense. So the feeling is: you’ve found like one of the most profitable leavers within the customer journey, where you can help the customer with a problem that’s caused them not to buy, and help the retailer to complete the sale, and you want to focus on that rather than just being a tool that people can do whatever they want with SMS?
Dennis: Yep. As Attentive and Postscript scale, we do have a lot of mutual customers. And our feeling is that they are ‘platforms’ and that’s great, but we are a product that’ll be hopefully integrated and built on top of postscript and attentive as like a premium managed service because they’re not building that they don’t want to do. They want to build the massive enterprise scale billion dollar business. And we’re okay. Building a $10, $25 million a year business. So I think that our, our ambitions are a little different. I think longterm, we have a better chance of getting to where we want to go and having an outcome that’s financially happy for myself and my co-founder and whoever potentially buys us, or we could just run the business for a long time and be happy with with where we are. So I guess my ambitions aren’t as big as the platforms.
Dan: I guess if reaching a life-changing amount of money is the goal, if you’d been funded from the start, you’d have to grow much larger to reach the same result, and perhaps give up lots of control. And I guess now that you’ve made it to this stage you’ve put yourself in a luxury position where you can choose what you want to do next?
Dennis: Exactly. A lot of VCs have been emailing us for the last six months. Like, Hey, are you trying to be the next PO MailChimp of SMS? We’d love to like potentially lead around and invest in you. And we’re like, no, we just want to build a 10 million plus dollar annual recurring revenue business that’s profitable. And they’re like, Oh, that’s very refreshing to hear you just want to build a lifestyle business. I’m like, it’s a pretty good lifestyle to me. I don’t know. It sounds great to me. Yeah, exactly.
Dan: Funny – so they’re sort of ‘negging’ you?
Dennis: I know it’s almost like “we want you to think bigger. You’re too small, but that’s cute that you want to do that.” – I’m like, “whatever”.
Dan: It sounds like you’re being sensible to me, especially when you’re growing well as you are and adding new platforms and features. You talked a little bit about A/B testing for example – that must mean you’re interested in driving good results for retailers. In terms of that, what works for ecommerce sites you’re dealing with at the moment. Is it the coupons/voucher codes that work the best or are there other approaches that people take?
Dennis: That’s a good question. And it’s also part of the roadmap that we’re working on right now.
So historically – and I mean like the last two and a half years – we’ve only led with discounts because we usually send an initial message that says something along the lines of like, “Hey Dan, this is Dennis from LiveRecover. I saw you were checking out our abandoned cart recovery tool for SMS. Great choice. Would you be willing to see if I can get you a good discount?” – if I asked you that, the reply rates are so much higher than if I say something like “Hey, Dan saw you were checking this out, did you need any help?” – It’s like… off. I don’t need any help. But if you ask me for a discount, I’m going to say, yeah, I want a discount. I also need help.
Also the initial first impression of our app, if we didn’t do the discount offering, was ”Well, hang on your homepage, you advertise LiveRecover can win back 20% of abandoned carts on average” – and that’s true if you lead with a discount. If you don’t, your first impression is going to be, “Wow, your app sucks because I’m getting a lower reply rate or a lower recovery rate than advertised on your home page”.
We didn’t want that experience to be the way people had their first impression of our app. But now we do realize there are brands that are like non discount centric and say, “Hey, Dennis, we run an apparel brand”. Or maybe,let’s just say in an ideal world, Louis Vuitton said “We want to use LiveRecover, but we don’t do discounts. Can we use our app?” the answer would have been, no, sorry. It doesn’t work for our app.
But now we’re adding something called message segmentation. So you could say if the product equals this product ID or product sku, then don’t offer a discount, but still send a message, right? If the average order value is less than say $40, don’t send a discount text, send a “win-back” text, or a “discovery” text, or an “educational” text. We’ll have some kind name for the journey or type of template that we ended up calling that, that questioning or formatting of that quote of that initial text.
That is going to be how we grow the product to not only lead with discounts, but I think that’ll still be the main goal of most of the conversations – to win someone back. Especially in 2020 and 2021, there’s just so many direct to consumer brands that are trying to fight for ad space on Facebook and Instagram and everywhere that you kind of need, you’re willing to overpay, to acquire the customer the first time, because given that you’re not selling like a mattress where the repeat purchase rates pretty low. If you’re selling a consumer packaged goods product or a product that could be recurring, do whatever it takes to get that first sale, in my opinion.
Dan: That makes sense. And on recurring products, or sites with loyal customers where people buy again and again, do you have features to segment them?
Dennis: So you could say: first-time customers are offered 10%; returning offered nothing, or if we’re turning off or 5%, or if the customer is spending a cart value of greater than a hundred and they are a VIP customer.
And what I mean by VIP might be something like they have a total spend of $500 plus, which might be something that’s applicable to like a clothing brand like FashionNova or anyone that has high AOV and just total lifetime value of a customer. You can definitely do all that kind of stuff, the same way that if you have sale items, you could exclude items because of the sku and product ID, or you could do it by tag. I wouldn’t say it’s super advanced, but we have all the options.
Dan: That makes sense. And that applies equally with the BigCommerce integration and woocommerce?
I know some of them are different per platform because the platforms each have different rules around what order details we can get on the customer. That experience I was just talking about was specific to Shopify.
Dan: I guess the vast majority of your customers are still on Shopify?
dennis: We are working on growing that, right. That’s why we’re hiring people to help with sales and business development, but, WooCommerce is new for us. We haven’t done a lot of content marketing. So I think if you search “LiveRecover WooCommerce”, you probably get taken to our WordPress profile page, not even our website, and we need to change that. Soif anybody’s out there in a position that thinks that they could help us: we’re hiring, hit us up.
Or if you’re a merchant we’d love to, we’d love to talk to more WooCommerce or BigCommerce merchants. Because it’s not as public as a community as, as Shopify. For example on Twitter, Shopify is what everyone that’s got an audience talks about. WooCommerce seems to be like, what most of the developers are using that aren’t very vocal.
Dan: Kind of funny because there’re far more installations of WooCommerce, but I suppose many of those are blogs or content sites with a small ecommerce add-on, rather than specifically ecommerce businesses.
Yeah. 3 million installs. I’d be curious what the actual volume though, like the gross merchandise volume is, every year from WooCommerce versus Shopify. I’d imagine that Shopify is – even though they’re half the size – they probably do more.
Dan: I don’t know the answer but I’d very much expect so. Have you done much with BigCommerce?
I met them at a dinner with @web actually, at 2PM dinner – the newsletter 2PM. That’s how I initially got into touch with them. They’re in Austin – local to me. And so is CrateJoy, which is why we integrated with both of them.
Dan: And how is it going integrating with BigCommerce merchants?
I think our app is still in the “pending approval” stage for BigCommerce right now. So once we’re approved, I hope that they’ll give us some love and blog about us, but we will definitely be blogging about them, or just to hopefully get some inclusion because they have a product that people use.
Dan: Lots of BigCommerce merchants seem to be a little higher revenue than Shopify, and Shopify higher than WooCommerce generally.
Yeah. And I mean, Magento’s huge too, right?
Dan: Yeah. And the Magento developer community is huge too – I think the number they often quote is a developer pool of 300,000 globally.
Yeah. That’s what I mean. We have a Magento plugin also, but I think we have one store. We have one Magento store. So if you know anyone who’d be interested, or if anyone reading this wants to be number two and number three, get in touch.
We’re definitely trying to expand into other platforms’ customer groups. For example we’ve started to talk to a WordPress agency that works with all these top WooCommerce stores. I think they’re based here in Austin.
Dan: Great, so we’ve talked about the product, and what you’re trying to do, and where you’re at, and how you’re marketing it. What about results – are you able to share some of the detail on metrics you help retailers achieve? For example, what percentage of SMS you send out get a response?
Dennis: Um, I would say it depends on the brand, but on average, it’s over 50% – I think it’s 55% platform wide. Some stores are down below that, the lowest being I’d say 45%ish and some being as high as 65%. So really high reply rates, but that’s because of the comment that I mentioned earlier about like leading with a discount – that obviously inspires a reply more so than, than not. That’s not necessarily the final outcome – a reply rate means “yes” or “no”. That doesn’t necessarily mean they’re all “Yes, I want to engage”. But like you would assume that 99% of them are. Yes. If they’re not interested, they usually don’t respond. Some people say “no”, but usually not.
Dan: That’s really interesting. And what would you expect that would be without the discount?
Dennis: I mean, again, it’s different for different brands, and different depending on the verbiage. These are some of the things we’re going to start testing more too, like “what if you have an emoji in the conversation?” vs what if you don’t; “what if you have a question mark?”, “what if you make a comment vs a question?”, ie what gets the best reply. I think before when we were testing with no discount, it was as low as 30%, but that doesn’t necessarily mean that it couldn’t be higher if you hadn’t better verbiage.
I don’t know exactly what we were testing with. Something like “Could I help you?” versus making a joke and maybe lightening the mood and being playful, might be better, right? There’s a lot of ways to optimize, but we haven’t gotten that far yet in terms of testing.
Once we start rolling out message segmentation and we can do analytics on reply rates, per message and her offer and things like that, our analytics product will get a little more sophisticated – it will be fun.
Dan: One thing you said there, and mentioned a little earlier, is that it differs between different brands. Are there some kind of sweet spots there? What would be a good brand and what would work less well?
Dennis: I’ll make an example between a couple:
- A non-luxury apparel brand: they have about an average order value of a hundred dollars because they sell typically hoodies or sweatpants or sets or whatever joggers.
- And then there’s like a makeup brand. Let’s call it $15 to $30 items.
The brand that’s $100+ – if they give away free shipping or even 5% off – they’re most likely going to have a way better recovery rate, than the brand selling lower priced items and giving 10% off even, because the perceived value of that discount is better because of the price range.
So it’s like the price and the AOV usually dictates. A lot of our higher average order value products get much better results by giving away less. That’s all just because of the perceived value I think. That’s not to say that stores with a lower average order value can’t make a great offer.
A lot has to do also with the brand’s coupon strategy. If you’re a brand that doesn’t give away discounts ever, or at least very little, and you maybe only do it on your welcome popup or first purchase, and you don’t really run retargeting ads, and then you offer something like 5-10% off or free shipping via SMS, it’s probably going to convert very well. 20% conversion maybe, or higher.
But if you’re an apparel brand and you always have 25% off sale fixed at the top of your site, and you always run 30% off retargeting ads, and then you offer that same 25-30% off to try and win them back via SMS, you’re going to have pretty poor results because customers are already conditioned that you are a discount-centric brand. If you’re not providing the customer anything more than they were already given elsewhere, it’s not going to do much.
Some customers in that scenario might respond and say, well, yes, I want a discount. You give them 25 or 30% off. And they say, well, I already had 30% off. “You’re not giving me a discount at all. Why’d you text me?” So in that case, you have to be pretty conscious and, maybe you want to create segmentation roles. If they already have a discount on their cart, create a bigger one. If the value is 25, maybe 30, if it’s 30, make it 35. There are things like that you can do to keep up some of the performance.
If at the total other end of the scale you’re not a discount centric brand at all, you’re going to have great results even giving away things like free shipping. If you very rarely give discounts, even if you already have a free shipping offer, just even reiterating that may work well, because people don’t think of you as a discount centric brand.
And it also depends on the type of purchase. If you’re selling just an impulse item, like a watermelon slicer and avocado peeling tool, that’s a viral gag gift on Facebook, right? You’re not going to have people who care as much to get a discount on that as you are if they’re buying on your brand and they’ve come through organic or paid search.
So I think there are a few factors, those things definitely determine the performance.
Dan: Makes sense. And for conversion, do you measure it from voucher code use or do you measure it from literally comparing who you’d sent messages to prior to them completing a purchase?
Both. So we have attribution models, and two ways we work within that are:
Firstly – we use unique discount codes by default – which are a single one-time use – one code per customer. So if they use our code, then we obviously know that it was generated by us – that’s been from our efforts.
Secondly – now in a scenario where I offer you, Dan, a discount over text, you say, “yeah, great. I want the discount”. I gave you 15% off, but it was the scenario I just mentioned where a brand did not give us the best offer. You say, “Oh, I already have 20% off”. You buy it from the link that we sent you back to checkout, but you had 20% off. Why would you use our coupon? You’re not going to, you’re going to use the link that we sent you back to checkout. So in that scenario, we tag the URL and say if you went back to checkout on a last click attribution basis, if you purchased from our link, regardless of the coupon used or no coupon, that’s attributed to our efforts.
So those are the two ways.
But now imagine you did not use our coupon code, and you did not click our link back to checkout, we wouldn’t take credit for that. A lot of our competitors do implied attribution revenue. Meaning if I had a conversation with you, and you just bought within the next seven days, they assume that was because of them. Even if they text messaged you, even if you didn’t respond, if they texted you and you bought within the next seven days, they count that as because of them.
Dan: A little like Facebook view attribution.
Dennis: Yeah. It’s like click versus view attribution. We did not do that: it’s only on click. It’s only on click for the link or coupon code for codes.
Dan: That makes sense. And, talking about attribution and measurement: You talked about putting some advanced analytics stuff together. What’s in the dashboard at the moment? How do people understand what value you’ve delivered?
Dennis: Let me see, I’ll give you the exact numbers or like what we give.
So the high level metrics on your dashboard will show you things like:
- How many conversations did we have?
- How many orders did we recover?
- Total sales recovered
- Average order value of sales we recovered…
- … and then the recovery rate
So those are the main dashboard analytics, but then obviously you have like a dashboard or a date selector. So you could look at custom dates this week, today, this month, whatever.
Then you have visibility.
I’m looking at the dashboard now – let me give you some numbers. One store today we’ve recovered $3,250 for them yesterday, 6,000 orders total this year, almost a million dollars for this year recovered over texts with us.
Dan: That’s pretty great. That’s a good case study.
Dennis: Yeah, for sure. And if I’m them looking at the dashboard: You can jump into the conversation. You can see initial messages, can jump into ones that have been recovered, where they haven’t been recovered to try and get an idea of why, and various other things. Then the insights will give you things like reply rates. Yeah. This branch reply rate is 62.4%. And then the recovery rate 18.5% for the last 30 days, segmentation…
Dan: 18% is as a percentage of the overall not the percentage of the replies. Right?
Yeah, exactly. So it’s actually, yeah. Yeah. So, I mean, I guess the recovery rate would be, I mean, I don’t know, it’s 17 from 58. That’s like, yeah.
Dan: It’s like 30. Yeah. 30%.
So with the dashboard we do the high level metrics.
Then there’s a “products” report that gives you an overview per sku for total count per sku.
And then “revenue count contextual” is turned off right now, but this is like text analytics. So if you want to not have to read through all the conversations for the last 30 days, you could just look at that and it extracts the keywords that are used the most.
Once AB tests are set up, you can see the different values of AB tests. So you could see their base levels, 5%, you could make a fix like add a free shipping value, and can test the difference.
And there’s segmentation
Dan: Yeah. That makes sense. I like that as well, so you can use the dashboard to optimise performance, like a dashboard is supposed to allow, rather than just playing back numbers you can’t impact.
Dennis: Exactly. So you can make all these different offers, or add something like an end of year sale, or you could just choose to keep it running. Yeah. It’s been a cool journey so far.
Dan: And with the Shopify thing, do you have any worries about them cutting you off or anything like that?
Dennis: I think Shopify is really generous with working with app developers in their ecosystem and like they gave recharge and all these guys, like a year to prepare for their migrating, cutting off third party checkouts.
And so I think now that their upsell API and their subscriptions API are public and the new checkouts here, I think the whole ecosystem of apps actually built on top of Shopify make more revenue every year than Shopify makes in revenue for themselves through the platform.
So I think they know the value of what they have on top of them, but you’re right. I mean, it makes sense for them to acquire apps that build things that service a lot of their customers. I don’t think that Shopify will build what we’re building, because what we’re doing is not a potentially billion dollar a year business to do it.
You know, we have to have how many texts and agents – that’s a lot to do that scale. And we’re happy building a $10 to $20 million a year business, and that’s maybe not something that Shopify wants to do.
They’d maybe rather either go and say, “Hey, we want the MailChimp of SMS so we can just plug it in Shopify”. Or maybe they don’t because Shopify is a software product. They don’t want to deal with servicing customers really. Like you just use their stuff. And that’s what the apps are for. They build the software and then anything else that’s service related or support related pretty much comes through apps, and that’s allowed them to reflect a lot of the real support that they need. All they have to do is support the developers, and the developers support their customers.
And so because of that I have a feeling that they’re probably gonna, like not really ever kill the app ecosystem. They play pretty nicely with us.
That’s what I think long-term in general, I do think that Shopify will get into SMS though. I do think that they’re probably going to have basic SMS reminders or maybe even an SMS API on top of Shopify. And we would love to be a solution that gives you the best premium experience for that. Because most stores don’t have the resources to have a full time managed channel. Right? It takes how many shifts are there on a Monday through Sunday, right? Seven days a week, eight in the morning until 9:00 PM supporting three times zones. It’s like at least five people full-time, and most companies at that scale at that point, you’re probably pretty large if you have five floats on customer service reps.
Dan: Yeah. That makes good sense. And it’s also probably the kind of thing that would push more people towards you because people wouldn’t try out SMS a little bit.
Dennis: Exactly. The education is kind of like: more market pull towards SMS means more people will try chatbot versions, which means they’ll maybe try us too.
There’s a time and a place for automated SMS. Like creating a list and doing a big outbound blast to 50, to a hundred thousand customers. It makes sense first over email. But if you do it over SMS, you don’t need a human powered tool, like go automation. That makes more sense.
But maybe if they try automated versions they realise they have similar philosophies to us, like humanizing their messaging to get better results, the way that we are trying to do.
Dan: One thing you talked a little bit about there was agents. How is that difficult for you to scale? How did you manage that, how do you forecast how many people you’ll need? How does that side of things work?
Dennis: So the main metric that we use to measure on when we need to hire more reps is time to first response. And then time to reply. If our reply times are slowing down, that means that our reps are either overwhelmed or not doing their jobs. So we have to check in with them based on those numbers. This is a high level like, okay, if your time to reply and tickets responded to per hour is increasingly different than usual, um, are you away from the computer and not doing your job, or what’s up? Is it just because there’s so much volume, and we can tell because the other agents have a lower reply times, or that it’s individualized to one specific agent, right?
So that’s one way that we usually plan to scale, but our reps have like a really, really powerful chat CMS that we’ve built, which is kind of like the main tool that allows us to be so efficient and not have a 50 person texting agent team – because we built really powerful tools to basically allow them to send a cart send the coupons and the back to checkout URLs, fast replies to certain common messages about failed billing or shipping, or this and that.
And they don’t have to respond manually by typing every message. There are intelligent pre replies and buttons that create the coupon code with the Shopify API on the fly and create short URLs that redirect back. So one texting agent can manage hundreds of conversations an hour. They’re on a computer, responding to customers who are on their phones, and texting from a computer can be made to be much faster than a phone. So it’s really like what we’ve built on the back end that’s allowed us to be as efficient.
But throughout the next year, I imagine we’ll probably double the size of our texting agent team, which is about 10 right now.
Dan: That’s pretty great that you just have 10 people handling so many customers. Roughly how many, how many texts are you sending a day or a week or a month?
I’m not a hundred percent sure off the top of my head, but I mean… I think it’s millions of texts a month. I don’t know about that. Might be hundreds of thousands to millions of texts a month. Yeah. Something like that. Yeah. Definitely if I think about it. Millions. Yeah.
Dan: So automation and optimisation for your agents has saved a load of time internally for you, and that’s one of the reasons you’ve been able to grow so quickly. Anything else along those lines?
Dennis: Yeah, integrations: We integrate with Carthook, OneClickUpsell, Recharge, and Bold subscriptions… oh and now Gorgias – the customer service platform. The integration with Gorgias is useful because it means we can open tickets for customers. We don’t have to tell the customer “please email support@brand…”, we can just push a button to open a ticket, and that passes all of the text conversation straight to the brand, and the customer gets an email letting them know too. Zendesk and Zoho – we’ll do those next too.
Dan: It sounds like you have a great roadmap, and you’ve done so well so far, particularly as it’s been such a strange couple of years in the world. What do you think of what’s happened over the last year, and what’s likely to come in the next couple of years?
Dennis: Yeah. It’s been weird. I’m empathetic for people who are in poor situations, of course, because I’m not in a poor situation. And I only want people to be empathetic towards me, but it’s, it’s very like, COVID accelerated things that were not secret. It’s not like, “Oh my God, I have to take my business online”. Like it’s 2020. You could have, if this was 2006, and you said that I would have almost felt bad for you, but even in 2006, I was building websites and on my space and I’m like 15 or 16 or whatever, 17 at that point. So it’s like, okay, it’s not 1998. It’s not 2008. It’s 2020. So if your business is getting decay and your business is hurting, it’s probably something that would have happened in the next five years anyway.
It’s unfortunate that it had to be a pandemic and a disease, and it’s terrible that’s affecting people and killing people.
But for ecommerce: It’s like businesses that didn’t keep up, it’s almost like a refresh. That’s kind of inevitable, in my opinion. It does suck, but it’s not rocket science to think that you should have taken your business online, and that ecommerce was the future. It wasn’t a secret.
And there are thriving businesses that I see, like, I mean there are even some restaurants here locally in Texas, one called Torchy’s tacos. They added 20 new restaurants this year during COVID and they plan to add 40 more next year.
And people are like “the restaurant business is dying” and I’m like, well, these guys just raised $700 million and they’re making two and a half million per restaurant. And they added 20 this year and they don’t have a drive-through you have to go in there and sit down”. So like, that’s, that’s the opposite of what makes sense right now – you have to sit down, and you have to get their deliveries through their own system – it’s not Uber eats or distributed through all these other companies that take a massive cut in margin.
So I’m like, I don’t know, I do feel empathetic, but I also see people winning and losing on both sides. And I’m like, if you didn’t make it in 2020, I feel bad. But maybe there’s a time to like look forward and rethink some things.
Dan: So you mean that you think all this change also throws up opportunities that would usually feel counterintuitive?
Dennis: I mean, I read about some kid through Twitter. He was like, he built a COVID directory that pulled in all the APIs from the CDC and population data and whatever he got. He put Google adsense, and native ads like Taboola, or Outbrain and some affiliate offers, he made like five or $6 million off a COVID map website because of the amount of traffic. Right? And it’s like, Oh, it went number one on Reddit or hacker news. And then it was indexed in Google and it showed up on page one for like a couple of weeks. Now it’s like, not even page one at all, becauseit’s a little bit more controlled by Google and official sources and whatever, but I’m like, wow. Like the fact that like a kid can pop up and make a life changing, a lottery winning amount of money, building a retirement fund. If it’s a week, I spent a week on that or a weekend tying in APIs and libraries that he didn’t even write.
It sucks that there are so many negatives, but there are also big opportunities.
Dan: Thanks so much, Dennis. That’s all super useful, and I think people will have learned a lot from reading this.