In today's episode, learn how to automate your support. Join Steve Weiss as he discusses with Romain Lapeyre, from Gorgias, on how companies can optimize their customer service and chat experiences for their customers.
Episode Transcript — Automating Your Help Desk w/ Romain Lapeyre of Gorgias
00:01 Speaker 1: You're listening to The Spend $10K a Day podcast. Brought to you by the performance marketing experts at MuteSix. This is your source for cutting edge insight into the world of online advertising. From the team with more Facebook case studies than any other agency on the planet. Here are your hosts, Steve Weiss and Stewart Anderson.
00:24 Steve Weiss: Hey, everyone. Welcome back to another day in the paradise of e-commerce, at the Spend $10K a Day Podcast. Today, we have an incredible guest with us today. We have Romain Lapeyre. Romain has an incredible startup in the Shopify ecosystem. He's a all in one help desk. He's solving an incredible pain point for a lot of e-commerce stores, and I'm excited to have him on our podcast and really dig into the weeds of how companies can really optimize all their customer service and chat experience. Well, thank you so much, Romain, for coming on our podcast. I really appreciate it.
01:03 Romain Lapeyre: Sure, thanks for having me on the show.
01:04 SW: So, first off, before we jump in, tell me a little more about yourself. How'd you get into the Shopify e-commerce ecosystem? Tell me some stuff you're passionate about. I just wanna get to know you and I want our audience to really get to know you as well.
01:18 RL: Sure. It's a story with multiple steps basically. I'm kinda passionate about automating stuff and saving time so, that's how we started with my co-founder, with Gorgias. Initially it was a Chrome extension to write e-mails faster. And so, we started that three years ago and we continued now to have 80,000 users on this product. And so basically...
01:44 SW: Wait, wait. So first Gorgias was a chrome extension to help people write emails faster, correct? That's really important.
01:52 RL: Yeah, exactly.
01:54 SW: And today the product has pivoted. So, you saw a need in the market and you pivoted the product. Tell me where the product is now.
02:04 RL: Yeah. So, actually, so, we pivoted from this chrome extension to doing some machine learning in customer service. So, basically the biggest users of the Chrome extension were support agents. So, we basically sat down with them and watched them do support and we learned that half of support questions are the same. So, we basically wanted to find a way to suggest responses to common requests basically. And so, this was pretty interesting. It was in 2015. So, before machine learning was height basically and so, we basically noticed that it was pretty hard to do that from one campaign to another so, we decided to focus on the Shopify space and to be able to help desk from scratch.
02:55 SW: Interesting. So, the bigger picture is, AI machine learning has evolved, I'm sure you'd agree over the last couple years, and instead of paying a significant amount of money to have customer support agents, whether they're here in the States or overseas, the main solution that a lot of stores and e-commerce brands are having is to automate their support, correct?
03:15 RL: Yeah. So, it's actually a combination of both. So, it should take an average, like Shopify store basically. What would happen is that they used to have some agents in the US and since the margins are not always very high, what you wanna do is to optimize the customer support and so, then you hire some support staff overseas, that's a bit cheaper. But like you pointed out there is an opportunity with the automation basically to have your staff in the US and to not outsource your support. But instead of that, instead of responding with canned responses all the time, what you can do is set up some automation to respond faster to customers basically.
04:00 SW: Cool. And just from an automation perspective, why is automation so important right now? Why do automation versus having a chat sales rep or a chat help desk.
04:15 RL: Yeah, so, it's actually a pretty interesting time for automation because last year so, in 2017 we've seen, there's been a big rise of messaging, specifically for an e-commerce so, think of tools like many chat points and so, Shop Message that send marketing messages to people on messenger. And so, when you are a merchant that uses a marketing automation on messenger, you get this big spike of conversations from customers because of this marketing automation. And so, the challenge is, "Hey, how are you gonna deal with this volume without hiring another five people to deal with your support. And so, this is why automation can be helpful because it basically comes with the rise of conversational commerce.
05:05 SW: Interesting. So, pretty much what you're saying is that there's a lot of messages. [chuckle] Some of the messages are probably repetitive. And the number of messages is rising and if brands and e-commerce site owners aren't thinking very logically about how to take care of all these requests, they can be inundated and all these requests can fall through and slip through the cracks. So, even if you have your own chat agents, you should also consider having an automation platform as well, so you don't have to always be on the ball, correct?
05:41 RL: Yeah. And also... So, if you fake the support messages that a merchant would get, a big chunk of it is repetitive. So, it's more interesting for the agents to focus on the complex cases or to basically spend a lot of time personalizing the experience for the customer and the responses rather that just say, "Hey, here is your tracking number."
06:07 SW: So, how does Gorgias work? Obviously, each brand has their own messaging of customer support. They have their own products. How do you set up Gorgias? Is it hard? If I'm a store owner, how do I integrate this into my stack?
06:25 RL: Sure. So, it's actually pretty straight forward. It's a matter of connecting Shopify, connecting your emails, connecting your Facebook page. And from there, once you've done that, you start getting your support messages inside Gorgias and you start seeing information about customers next to those. So, for instance you would say, let's say Steve you sent a message about delivery status so, that the agent would see the message. And then next to it they would see the tracking number, delivery status from USPS. So, that's what the platform looks like. So, it helps agents be more efficient basically. And then there's this automation component which is sort of and ifs this and thats thing. And so, you can do something like if somebody says, "What's the status of my delivery?" then check if there is an order for this customer, and if there's a tracking number, automatically respond with this.
07:24 SW: Cool. That's awesome! Tell me this, what type of brands should automate and when not to automate? Who should have automated support? Obviously, large brands, small brands. Who, in your own words; we do a lot of acquisition work. This Podcasts is all about performance marketing. How do I drive performance? How do I drive performance? How do I stop retention? Obviously, doing great work on the customer service side as well is gonna increase retention, but talk a little more about which types of brands should have an automated support.
08:01 RL: Sure. So, it's kinda of a functions of the lifetime value of your customer. For example, if you're a luxury brand, what you would do is, you would hire some agents to give a personalized experience to everybody regardless of what's going on. But, the thing is if your average order size is $10 then things are completely different because the resources that you can spend on customer service are much more limited. So, this is where automation is helpful. So, I would say most of the merchants would do that except the very luxury merchants.
08:38 SW: Interesting. What companies; give me examples if you could, of just a couple of companies that are super successful with automation from a customer service perspective.
08:49 RL: Sure. So, of course we mostly work with companies on Shopify Plus. For instance, we work with the company in Toronto called "DD's", they are drop shippers and sell sportswear's apparel. And so basically, they have this challenge of having... A year ago, their support was email based and they were not investing a lot in it. And so if you were to contact the support, you had to wait for one or two days to get a response. And so, these guys couldn't invest a lot in customer service because of the margins and so, they decided to invest a bit more in automation. And at the same time, what was interesting is that there was the transition to messaging. And so, they're pretty successful with it because, so, they have this whole marketing on Messenger that generates a lot of conversation. And I think they get 10 very common request that ranges for delivery or returns, or refunds, things like that. And this is automated. So, drop shippers on Shopify plus is a big one. And then you have all sorts of brands like bracelets or jewelry, things like that.
10:01 SW: Got it. Interesting, interesting. Some of those move over to automation, what are some of the risks? What are some of the pros and cons of having full bot versus help desk automation?
10:18 RL: So, basically, it's a...
10:21 SW: Or point the difference between a bot and a help desk. I think a lot of our readers as they're building... Or listeners, should I say as they're building e-commerce site, they're trying to understand, should they have a bot or should they have a help desk and or they should have person. There's a big differentiation between all three.
10:39 RL: Yeah. So, it's typically, the market is moving toward a combination of the two. So, combination of automation and then having agents responding to support requests. And so, the automation piece can be broken down through a bot that you would have on messenger. A bit like surveys; you ask a question and the bot say, "Hey, how can I help you?" And you would have a drop down saying, "With my return.", "With my delivery." And you click on something and then that triggers another response from the the bot. So, bots are pretty good for messenger because messenger makes it pretty easy for bot companies to implement those and they are pushing for this behavior. So, that's pretty great. On our side at Gorgias, the approach that we take is to put the automation directly inside the help desk. And so, there are also a couple of advantages here. So, one is that you can have automation on any channel. And for a merchant periodically who get 70% of their support is email. So, having automation on email is pretty valuable. It's cool to have some automation that works on email, messenger, chat. So, that's why it can be helpful to have it in your help desk.
11:50 RL: And another cool thing about the help desk is that, at least for Gorgias we connect with your Shopify accounts; we connect with ship station; we connect with loyalty points. And so, you can combine this automation with what's going on with the customer. So, a simple example would be, if a customer has more than a 1000 points then, routed to an Asian guy. If not, check what the question is and automatic your response to it.
12:20 SW: Interesting. My last question is, I was just trying to figure out, you're an e-Commerce store, you're trying to figure out what you should be doing on the bot side, what advice would you give to new business owners? Who were considering making a solution or making an investment? When do they make an investment into chat?
12:48 RL: Yep. So, I think it's actually a good time for most of the business owners to start making an investment in our conversational commerce so, marketing on messenger, support on messenger because basically it gives them an opportunity to create a relationship with customers. So, instead of just sending impersonal responses, you can send messages in the inbox of the customers on the messenger app, which feels very intimate. I would recommend to do it now and other helpful, I think, is to approach it as Email Marketing basically. So, when you send an Email campaign, what you wanna do is to make it super personal. So, you want to send the right message, the right person at the right time. So, you need to do something similar with automated customer service. Meaning, if somebody asks you a question, you want to tell them, "Hey, thanks for being a customer for your," maybe use the order number, use the variables, like the location of the customer. Because what's valuable for you is to show that your message is specific to the customer. And so, after talking to you, the customer needs to remember something positive about the experience. So, actually a pretty good example is a company called BarkBox. They do a subscription business for dogs. So, it's like toys, things like that.
14:16 RL: And so, the entire brand tone is around dogs. So, if you contact support; let's say on messenger you would get a gif of a dog playing with something and then the language that they're going to use in the responses is gonna be doggy. It's a fun experience. So, when you interact with support, that's what you remember.
14:38 SW: Yeah. It's very native built that the support is very similar to the product. The tone of the support is that they're talking to someone intelligent. That makes a lot of sense. And I think the scary part of having automation or bot is the customization is, "What if a consumer asks a question they can't answer?" Like a chatbot. And I think from a store, that's something that I'm really thinking about is, "What happens, what if I ask a question that they can't answer?"
15:10 RL: Yeah. So, that's actually a pretty interesting question that everybody's thinking of right now. So, I think Facebook Messenger is focusing on it as well. The key component to this is the hand over part. So, having the bots connect really well to the agents and say, "Okay, at this point, I'm transferring you to an agent." And then, the customer knows that they're gonna have a human responding to this. So, right now, it's kind of an unclear area so, both are trying to do it. It's not working that well. But I think that's something that's going to improve in the coming months.
15:45 SW: I guess that's a big risk, is if they ask a question that just can't be answered. So, interesting.
15:50 RL: Yeah. So, the bot needs to know that they cannot answer and transfer you to a human on the automation.
15:57 SW: Cool. As we wrap up, I think you've really explained your product in a very concise way and really explained a lot of pain points in the chatbot industry. What would you say, just in a closing statement if you're giving advice to an e-commerce store that actually has some history, they just wanna improve. There's two different types of customers. There's the new customers who's starting an e-commerce brand and then there's someone, a brand that existed; what would you recommend is a good advice for both types of brands, obviously the new brand and also the brand that's been in the industry for a while, has experience, if they wanna improve the experience of customer support overall?
16:44 RL: On support specifically?
16:46 SW: Yup. On support.
16:50 RL: So, we work with a company called Keto-Mojo. They sell products to monitor sugar rates in your blood. And so, what's pretty interesting about them is that if you check the reviews on the website, 10% of the reviews are mentioning the quality of support. That's something we've also seen with all the stores, that they have this big focus on the quality of support and it starts being a competitive advantage. And I think using tools like automation, going through more personalization, having a brand tone, using messaging instead of classic email or phone helps using support as a competitive advantage. So, what I would recommend for these stores is to not say, "Hey, I need to hire some agents and provide some support because otherwise, my customers are not gonna be happy." More think of it as a competitive advantage over all the stores and try to have a pretty specific strategy here to differentiate.
18:00 SW: Got it. Any final thoughts before we wrap things up? I think there's a lot of information here. I think you've really built an interesting product. And I haven't heard personally of anything that's similar to this type of experience when it comes to building all-in-one help desk solutions. But before we wrap up, is there any other final thoughts?
18:21 RL: I think something that's pretty exciting is that there is a... Before there was marketing and then there was reports and now that the two can talk to one another you have the opportunity to build a continuous experience for the customer. So, they get some marketing emails, then they engage with support on the specific topic, and then you can adapt the marketing that you're gonna send them afterwards based on the support conversation. So, I think it's really interesting to interconnect those tools together and to build experiences for the customer, that are better. So, I'm expecting some changes in this direction in the coming years. So, I'd recommend keeping an eye on that.
19:01 SW: One other question, you sparked another idea, how do you integrate, that gentle integration of chat support and marketing? Can you give us an example of how a brand has, and I know you mentioned BarkBox, but can you give me an example of how a brand has integrated the marketing and chat and support together?
19:21 RL: Sure. So, for instance there is a brand called Darn Good Yarn. And so, they sell a subscription for yarns. And they're doing something interesting because they have people who are complete fans of the brand. And so, what you wanna do is to leverage those people to promote your brand. And so, for instance when one of these people has a problem with supports, they would obviously fix the problem. And then the person would be specifically excited about the brand, say, "Hey, by the way, I really love your product," and maybe share some pictures. Most of the brands just say, "Thank you" and that's it. But what this company does is that you can put these people on let's say a specific marketing list and when you wanna do a big push on a new product that you're gonna launch and then say, "Hey, thanks a lot for being a promoter of the brand. And thanks for reaching out to support recently. We have this new product that we launched. It would be awesome if you could post a picture of you on Instagram with this hashtag and the product". And so, that's a great way to leverage the work of the agent who tried to build a good relationship with this customer for marketing afterwards.
20:32 SW: Got it. Okay, interesting. That's really, really good advice. I like that. And that's pretty much it on my side as far as questions. The bigger picture I have is, how do you continuously interconnect that user experience? How do you connect that support experience back to the market experience? How do you intelligibly do upsells? We didn't really talk about upsells, but do you have clients or people use Gorgias that do upsells intelligently? Any strategies that you could talk about that could, speaks to upsells of allowing chat support upsells in a automated way? Is there anything that you talk to there?
21:13 RL: So, not in an automated way, but there are some interesting things in support though. So, there are many apps that do a product recommendation on your store. And so, the thing is, sometimes you have people reaching out to support saying, "Hey, I'm thinking of ordering this product. Is it a good fit for me?" And so, the cool thing about having a support platform that's completely integrated with a loyalty program, the last orders of the customer, that shows you basically the whole picture, is that you get some understanding of what this customer likes. And so then you can say, "Hey", or, "Yes, that's a really good product for you, but by the way, I noticed that you built this recently so, you might also like this other product that pairs really well with the one that you wanted to purchase". So, using these humans' interactions not to send tracking numbers again, but to make smart recommendations that are personal, is actually a really good way to use support as a way to upsell.
22:16 SW: That's incredible. I think that's that's the future of upsell, is being able to integrate that with the chat. That's what I'm fascinated with. And that's what a lot of our listeners are fascinating with, is uncovering that intel in the support and applying that to your ads. Being able to look at what are the pain points of the user? 'Cause I always tell my team to really dig into the pain points of the customer, even in the purchase process, post-purchase process. And then really apply that to the ad. 'Cause if, if you could address the pain points in the initial sales process that you learn from the support, then you're gonna crush it. Would you agree?
22:58 RL: For sure. Yeah, completely, 100%.
23:00 SW: So one other question, then we're going to wrap things up. Any advice? You're a marketer and you really wanna learn more about your customers. Any advice, when it comes to digging into the chat and digging into the data and the conversations on support? I think that's key and I think finding tidbits of information there. Any advice just on what to look for and how to, the application of chat, of support, all the way back up funnel to marketing?
23:28 RL: Yeah. So, support. Most of the people who reach out to support are either visitors on the website having questions on your product, or people having a problem. So, basically, all the people who are happy about the experience, who are gonna leave you good reviews, they are, they typically don't reach out to support. So, support is an opportunity to improve the experience by fixing the problems that your customers have. The thing is, it's pretty challenging to get this data because typically you would just respond and then you, maybe you have a bunch of tags that you used. Say, okay, this problem was about delivery, this other problem was about a broken item, something like that. And so, some of the best teams that we have, among the customers are doing a fantastic job at identifying the 30 or 40 kinds of problems that customers can have. And sometimes, let's say Black Friday and there is a new sort of problem. And they identify it really fast. And so, then they share it to the team, "Hey, we have this list of 30 problems." And they are really good at typing the tickets, or the chat conversations. And so, this way, they can react pretty fast to new trends. Oh and so...
24:36 SW: That's really innovative. I like that. I like that being able to stay ahead of the trends. Wow, that's cool.
24:44 RL: So, one of the cool thing on that is automation can also identify new trends. Put the tags automatically and so, I think that's one of the directions in which we are heading to.
24:54 SW: Interesting. So, pretty much 9/10ths of the product is getting ahead of that before someone writes a better review. That's what I'm always thinking about. How do we get.
25:02 RL: Yep.
25:02 SW: Ahead of it. The game. Well, Romain I really appreciate you coming on our podcast; Spend $10K a Day, I think our audience is gonna really enjoy some of the content you shared. If someone wants to get a hold of you, or get a hold of someone for just some basic advice on anything related to help desk support, do you have an email? Or is there a best way to get a hold of you or someone on your team?
25:23 RL: Sure. My email is Romain R-O-M-A-I-N at Gorgias G-O-R-G-I-A-S.
25:34 SW: Cool. And guys, audience all around the world, this is another amazing episode of Spend $10K a Day Podcast. I really appreciate you coming on and listening. And until next time. Thank you so much, Romain. And we'll be in touch, man. I appreciate all the information you shared.
25:51 RL: Likewise. Thanks for having me.