In-House vs. Outsourcing

Startup for Startup

Hey everyone, I’m Dani Lester. Thanks for joining the ride.

The deliberation between keeping work in-house and outsourcing it to a company or a freelancer is one of those deliberations that almost every startup has to face, at one point or another. Preparing for this episode, I was talking to a few people that gave me different angles on this important issue. I heard many opinions on the matter: some got burned pretty badly from outsourced vendors, whilst others said that outsourcing was an important step on their road to rapidly grow. Some of the people said that in-house is the definite way to create high-quality products, whilst their critics claimed that the in-house approach out of touch for the 21st century. How can you decide? You want the company to evolve, but what’s the right way to do it? What exactly do you give up on when you outsource? In what cases it might be worth it? And how can you decide what’s right for you?

In order to solve this dilemma, we talked to a few people. One of them was Roy Mann, Co-founder and CEO of monday. Since day one he’s been leading an in-house policy at monday. He has solid reasons for that.

07:00: I’ll give you an example. When we redesigned the logo, we could have gone to an external agency that would’ve done a big massive thing out of it and, have us work for hours picking stuff instead. And we had Mayan, which designed a lot of logos for ourselves. But when she showed me, like, I don’t know, like three, I said like, this looks great. Everyone said, yeah, it looks great. We picked that logo. And, and you know, it wasn’t like hours of, you know, talking about what the meaning of life is and why did we build the company? So that saves a lot of time.

But it’s not only about the fact that in-house might make things run faster. If you ask Roy, it’s also – and maybe mainly – about the team. The togetherness.

And yet, it seems like we’re witnessing the rise of freelancers in recent years. To better understand this rise we spoke with Kobi Masri, COO and CO-Founder of A.Team – a company that connects other companies to freelancers from different fields. A.Team doesn’t only connect though, it also creates teams. teams of freelancers.

01:15: What we do is we actually take top developers, designers, product managers, data scientists. We connect them into teams and then we connect them into clients.

Kobi claims that the demand for freelancers is higher than ever before: for every new full-time role, there are three open roles for freelancers.

06:20: The reasons for that are kind of a generation change. Our parents used to work in one company their whole life. My dad actually worked in two companies, and then he went into retirement. We change jobs every two, three years, and millennials, they actually don’t even work in one place.

But this generation change was not our first consideration when making the decision to do things in-house. What was more of a concern to us were the challenges an in-house policy creates, and there’s one challenge that is more prominent than the rest.

I think it’s mostly scale.

Vlad Mytskesky, a Senior R&D Team Lead, had to deal with this exact scale challenge: his team worked extremely hard, performed well, but there were still some projects they could never get to because there was always something more important to handle.

25:10: You want to scale your development abilities because for whatever reason, you cannot scale as fast as possible with your in-house development team or any other type of the team inside of your country, or for other reasons.

But scale is not the only reason to outsource. Another reason, which I’m sure you’ve been waiting for me to mention, is money. Deloitte’s global outsourcing survey showed that 59% of  companies outsource as a cost-cutting tool. 

18:05: And the last reason is flexibility. A lot of time, you’re very afraid to hire a full-time person, right? It’s a long commitment. It’s very hard to file. With A.Team, or with using freelancers in general, you can start working with the person. If you don’t like them, stopping the same second. There are not much consequences. You don’t need to fire anyone. You just stop spending time with them and you don’t pay them anymore.

30:00: But again, if you’re building the core of your product, I still think it’s better to do in-house because this is something that will last with you for a long, long time.)

So you want to grow, you want the company to evolve, but what’s the right way to do it? How can you decide what’s right for you?

Monday’s in-house approach did force us to face the cold hard facts: growing fast is making it harder to keep everything in-house. Vlad felt it, but his experience with vendors before he joined monday was not so good, to say the least, so he didn’t rush to outsource just yet. First, he wanted to understand what he wouldn’t outsource under any circumstances.

06:30: in our case, for example, the first line that we put is, if we’re building something for our customers, as part of the product, it’s something that we want to keep in house.

Saying that, Vlad refers to services we use only internally. There are many apps and platforms that monday’s employees use on a daily basis, but the customers have nothing to do with it. 

07:15: For example, we are using the tool called Lessonly for doing all of our certification process with partners, and the process is happening in two places. So there are some things where managing inside of Monday platform and some things we are managing inside of the Lessonly platform.

vlad (07:39):

And there are a lot of manual work to syncing up these things together. 

Vlad has found a project that fits outsourcing: his organic team won’t get to it, it’s not customer-facing, and it doesn’t “leak” private customers’ information to an outside vendor. But even if the circumstances are perfect, there are still so many things that can go wrong. The first is, of course, the choice of the vendor.

18:41: first of all, I think that starting from the first meeting, like you need to understand the values and the vibe of this company to understand that they’re on the same side as you, because there are different types of companies and different types of partners. And you need to understand from the beginning that the way how they approach work is similar to the way how you see things and see the world. And it’s totally fine if someone is not seeing the things the same that you are. 

Laura Levy, who leads an outsourcing project in monday’s customer success group, finds it important to simply learn from other people’s experiences: she asked people from other startups what they learned from outsourcing CS activity.

13:39: I had an opportunity to network with, um, a few people across a few different organizations who are in a similar position to myself to understand really, you know, what are the things that I need to look out for, or that I need to prepare for.

19:45: But at the end of the day, you just need to try. 

20:17: So I think you should just start small and see how it goes. And if it doesn’t go well, it’s okay. You just need to find the first projects, which are not so critical that you can afford yourself to lose.

Part of the challenge of choosing a vendor is the deliberation between vendors from western countries and vendors from developing countries. It’s cheaper, and therefore tempting, and it’s popular not only in the engineering world but also in other sectors. Some of these sectors I find surprising, like the animation industry. An animated film I really loved as a child – Sinbad – is the first feature-length film created using motion capture technology. This technology allows the recording of human actors to be used to animate the film’s characters. Sinbad was the first to use this quite revolutionary technology at this scale, and they did it with 200 animators located in Madras, India.

This Sinbad fun fact made me realize that maybe I was wrong for thinking that outsourcing to developing countries might cause lousy results. Laura Levy actually thinks that there are many vendors in these countries that she can learn a lot from.

17:50: offshore companies can provide a lot of benefits to companies like ours, like Wix, like any of the other ones that are using them in the sense that their operation is very mature. They’ve been doing this for many, many years, across many, many programs and organizations with a lot of different business needs and they’ve kind of done it. So there’s no reason to reinvent the wheel. We can actually use that to our advantage, right, make us better.

And yet, there’s no denying that outsourcing to developing countries can rapidly deteriorate – either because of cultural differences, lack of specific skills, or so many other reasons some of you might run into. I recently talked with a bunch of founders that had some really bad experiences working with vendors from developing countries, so it made me wonder if the lower price was really worth it.

18:00: why would I take someone that is less good? You know, you want to have the best or they can do stuff in an hour that others takes them weeks. Okay. So it’s not about the hourly cost. It’s about the impact. Okay. So what is the cost of the impact? And I think that like with amazing people, it’s always less expensive.

If you DO go for it despite the risk, Kobi claims you better have an in-house person who’ll consistently check out the vendor. 

34:10: use at least one person, one developer that you really trust that can always control and monitor the quality of the code. And also, think about the cost, not only of what you’re going to pay on the next month or two, but also what you’re going to pay for maintenance, and for the long term, when you’re going to build on top of this infrastructure that you just developed.

28:30: I don’t think that’s cheaper country are worse, and you can find great partners and great development teams in cheaper countries in Eastern Europe and in India. And in other popular outsourcing destination. So it’s not about the geographical location. And if you will get the whole sourcing team in California, they will be for sure, amazing because they [inaudible 00:28:51]. No they cost [inaudible 00:28:52] just because they’re in California, there is not so much correlation between these two things.

After thinking it through with other companies who’ve already done this, Laura did decide to go for a vendor who’s not really next door. It took some time though. First, monday’s leadership considered other options that didn’t include classic outsourcing. One of the suggestions was an automatic CS system.

10:20: using, um, you know, AI or any other tool, a chat bot, for example, uh, to help us, you know, maybe answer some of those initial inquiries or initial, uh, tickets that are more basic, more, um, FAQ oriented, and being able to provide that service quickly, um, automated. 

11:05: but when it comes to customer experience and customer support, having a person understand and have that emotion and that empathy is invaluable.

Cool, ah? so why should we get a vendor with dozens of CS reps if we can have AI bots instead? These bots would take the easy-to-answer sort of tickets, and would allow the human reps to focus on the more complex ones.

Play button

https://www.youtube.com/watch?v=jz78fSnBG0s

Stop button.

Oh. I forgot to tell you. the two voices you just heard were created by an AI model. In this case, the AI had been told to create a conversation between two other AIs that were both clever, humorous, and intelligent.

Play button

“Video” continues

Impressive, ah? Or creepy… anyway, the thing is – like you probably experienced in your last interaction with a bot – it doesn’t work THAT well when it comes to communicating with humans. Microsoft realized that in 2016. Its AI bot, Tay, soon turned out to be racist. Taking that into account, it’s clear why we at monday decided not to use automatic bots as part of our CS efforts – at least at the time of recording this episode.

10:45: bottom line, we want a human on the other side

Eventually, both Laura and Vlad got to a point that it was almost inevitable to outsource a specific part of their activity. 

They had found a vendor, but it didn’t mean they could just start working. They needed to clarify, both to themselves and to the vendor, what was expected. Kobi, A.Team’s COO, goes through this stage on a regular basis.

12:30: We use the time and materials model. In this model, you don’t agree about what the developer will do exactly, but you agree on a hourly rate or daily rate.

Kobi (12:42):

And then you will be able to adapt and work in a flexible way with the team in order to build what you want. You will usually do it in shorter sprints, like one week, two weeks or four weeks and, in each sprint, you define some scope that you will do, and you will pay to the developers according to the specific amount of hours that they worked. And then, in the next sprint, you will do the same thing. You will define what are the scope for the next sprint, and you will pay according to that. So, the actual ability to pay for exactly the number of hours you need for every week or every month, give a lot of flexibility to startups.

Although the model Kobi just mentioned allows flexibility, it doesn’t really solve one of the main frustrations outsourcing comes with: the consistent need to explain to the vendor what you want. This is not something we deal with when we work in-house, simply because everyone knows the company, the product, the culture, the goal – everything. Working with a vendor is quite the opposite in that regard, as Roy explains.

02:15: it’s up to you to know what to tell them to do. You can’t harness, you can’t build a team unless you know exactly what everyone needs to do. And in our case, it’s not, it’s not like that. Like, you need people, creativity, you need them to work together. You can’t just like, it’s not, if everyone will do their list of tasks that we will succeed. They need to think and tackle problems, face them, learn together and improve. So you can’t do that if the team is not working together, if each one is separate in, a different in place. So then you have someone managing them, sending out tasks, telling each person separately, what to do. And then they are the mastermind. If they know what to do, then it will come together. If they don’t know, they don’t. And they work way harder just to explain themselves, because no one has the context.

Roy (03:06):

Okay. So I think working together in house with a lot of things and it can be the fact that we’re doing our own marketing ads okay. So where the, people who are doing it, know the challenges we have in marketing, they know the product and what to like, what is new. So they know everything and they work together. So it’s faster. They don’t need to communicate all the time. Outwardly all the things that saves time, it makes for way better products. And yeah it’s hard, cause sometimes you need expertise, that in I don’t know, editing or marketing or like craft-manship that you may not have in the company. 

To deal with the lack of cultural, historical and technical context of the vendor compared to the in house team, Vlad worked very hard to make sure that there were  no loose ends. He did that right before stepping up to his next outsourced project: the “Zapier” project. Zapier is a platform that integrates different web applications with one another, and Vlad’s goal was to implement it into monday. Unlike Lessonly, which is only used by monday’s employees and partners, Zapier is something customers actually use. In a major league project like this one, the alignment between the in-house team and the vendor becomes twice as crucial from the very beginning. This is when each and every person must understand what the goal is, who the stakeholders are, and who’s in charge of what. At monday, since we love football so much, we call it a “kickoff”. 

Whistle blows, crowd is cheering

12:55: my main goal here was to give as much as possible context, both historical context and give the feedbacks and voice of the customer, what are the things that we’re trying to solve and to build the good background for the team who is going to work on this, so later on they can be very independent.

The will to give all the information in the kickoff meeting might turn out to be a disadvantage. Laura, working with a CS vendor, realized that she couldn’t just throw tons of information at them right at the beginning and expect the vendor to deal with it.

27:10: we’ve learned over time that our onboarding, uh, was not really meeting that need. We were throwing everything at them at once, giving them four intense weeks of onboarding and then, you know, putting them in the queue and saying, good luck to ya.

22:20: I think that’s formalizing everything too much is a problem because it’s also the problem of trust, not formalizing anything. And again, just saying we need to make our Zapier integration better is also not good enough. And again, we’re coming to the same problem of finding the balance, how to give enough context on what we want to achieve on how we want to achieve, how to align also our quality standards and for this, you need to do constant contribute specifically in the beginning, just to bring again the alignment about what do you expect from this team to deliver.

27:54: So what happened with this group is they came in, um, in the middle of June and they were trained only on the topics that we would consider basic.

28:09: then a couple of weeks later, we gave them a chance to kind of explore that knowledge, apply that knowledge, answer our users live on those topics. Then we added to their knowledge with intermediate training.

28:32: then a couple of weeks later, we gave them a chance to kind of explore that knowledge, apply that knowledge, answer our users live on those topics. Then we added to their knowledge with intermediate training.

28:52: So they’re really pioneering the new onboarding structure and they will continue to do that. And then we will also adopt that internally as well.

Whistle blows, game ends

The kickoff is fantastic. Everybody’s perfectly aligned and knows what’s expected of them. Now they can go out there and do exactly what they’ve been told.

ERROR

Yeah, well, the thing is that Roy doesn’t want them to do exactly what they’ve been told. Not exactly. His perspective about it is taken from the times that he worked as a freelancer.

Deja Vu

Roy (07:53):

I was doing outsource work for people for a while as well in writing games and, and building games, computer games. And I can tell you that, like, it was frustrating cause they gave me projects that I thought I can do like other stuff better and like improve and, and they didn’t really want to listen. So I had to do stuff that they thought was good and I didn’t. We didn’t share the same goal. So it it’s very frustrating.

This is why Roy doesn’t want his freelancers to just execute. He knows how frustrating it might be – both for the vendor and the in-house team. His experience as a freelancer showed him how easy it was tokeep quiet and just do what he’d been told, and that’s not what he wants from his freelancers. He wants them to talk, debate, and be more than just hands on the keyboard who focus on making us, the customer, happy. A very detailed kickoff might prevent that, don’t you agree?

Well, it  definitely might, but it doesn’t have to. One of the approaches that allows the vendors to feel more part of the in-house team – and therefore debate and challenge the in-house team more – is the outstaff approach.

19:05: out-staff means, Hey, I will be together with you.

19:15: We add a team to your team to work together side by side. A lot of time, it’s like we give a team, together with the team lead. So, the team lead would actually manage the team, but will work with you as the founder, or as the CTO, to run the project. 

Vlad, our dear team lead, worked on “Zapier project” in this outstaff approach.

02:00: in out-staffing model, usually you’re – quote, unquote – renting a developer and working with the developer as part of your organic team. But in a lot of cases, companies, including us are finding some specific balance between these two approaches depending on the project, depending on the company outsourcing company that you’re working with. And there are advantages and disadvantages to both of them.

13:25: So here we really try to find the right balance. So from one side we wanted this team to be working inside of our office next to my team. So they will be fully next to us with a lot of sync meetings, with a lot of feedback loops and good level of support that they can just go to the next table and ask questions that they have. So this is more towards out staffing model. But from other side, we wanted to give them enough context and enough background, so they can take decisions by their own and build this as their own project, to own it and being fully responsible for this. So I think that’s finding this balance. It’s the main challenge of running these kinds of projects.

In my eyes, outstaff actually tries to bridge the gap between the offshore team and the in-house team. Another important part in bridging is making sure the offshore team feels like a significant and real part of the company. Laura does it with many different tools: she provides in-house mentors to the outsourced reps, they’re invited to some CS group’s events, and they have a bonus structure, alongside a growth path according to their performance.

31:50: getting that buy-in and getting that motivation is really, really important from the beginning. And we need to kind of keep doing that as we scale, which will be interesting.

By the way, if you want to learn more about incentivizing employees, you better head to our first episode: “Incentivizing sales reps”. Anyway, at the end of the day, your goal is very well-defined: you want your people – either in-house or outsourced or something between the two – to perform well. You want them not only to perform well at the moment, but also to keep it going for the long run, whether it’s CX reps that are going to do this job every day for a while, or developers who’s code will be part of your product for ages. This is, by the way, why Roy prefers to make things in-house mostly. He emphasizes that with…

An example

11:30: when we’ve done the first campaign with billboards in New York. So everyone told us, okay, you know, nothing about billboards and, and like those kind of stuff. And we said, okay, let’s do it anyway. And so we, we started by buying the media. We had no idea what we were going to do, but we knew we wanted New York and wanted to do it. And, and, and we wanted to do a campaign, but we had no idea what to do. So we were really stressed obviously, because the time was ticking and we had a deadline because we bought the, the end. They told us this date you need to bring in. And we ended up really stressed and consulted with a lot of people that we knew kind of, and some told us like a direction to take like, say hello, we are here.

Roy (12:28):

And not just like try and sell it too much or whatever. So that’s what we did. And everyone worked super hard and guess what? The first campaign was like, okay. But it gave us a lot of confidence. We did it, we knew. And we had one mistake that was like really great. Like we took a lot of billboards, but we took two subway stations and everyone told us about the subway stations. No one told us about anything, about the billboard. And we said, okay. Subway. And then we did the second campaign on the subway and it was, and we measured everything and it was 10 times better than the first one. And then we started getting confidence and we built a lot of knowledge around what [inaudible 00:13:15] to buy, when to buy where to buy and you know, and, and we improve. And now we have like an amazing team doing both. So it doesn’t necessarily mean you’ll do the best work on the first time, but you need to be willing to learn, to be able to do everything. You know.

Whilst I understand the long-term benefits of an in-house policy, Vlad and Laura’s case studies emphasize the need to be open to outsource or outstaff. But being open to outsource and outstaff is not enough, and neither is choosing the right vendor. You still need to measure if the vendor is working properly, and even more importantly – improving over time. Laura shared with me how she’s achieving this.

33:35: They have the same KPIs that we have internally. Um, so really holding them to the same standard that we would hold anyone to in this role. And, uh, they’re doing great. Uh, unfortunately we have been really focused, uh, particularly in quarter three on delighted customers. Uh, our overall CX department goal is 85%. Uh, they exceeded that, uh, in the past month, 87%. So definitely exceeding that goal, which is great. Um, and we’re globally pretty close to the goal as well on 84%. And, uh, the, the next thing is also productivity. Productivity is really, really important, right? We talked about, um, a balance of speed and quality. So productivity lends to the speed side. So ensuring that they are producing, they are answering tickets hour over hour during their shift, um, is really important.

Laura (34:29):

And, and having eyes on that is, is a big thing. And we’ve been able to actually increase our comment count, uh, more than double since they started. So, um, and even in the past three weeks, we’ve doubled by, by, or I’m sorry, we’ve increased by two comments per hour, which is really big and just goes to show again how quickly they respond to change and to our business need and our requests. And the final metric is, um, well, we talked about delighted customers, which is customer satisfaction. Uh, basically, uh, any user that’s gives us a nine or a 10 on their survey after they’ve had an interaction, uh, ticket email with us, uh, is considered a delighted customer.

To be honest, I’m sitting here in the studio and trying to figure out how to end this episode. Although the company I work for is in-house oriented, this episode really made me feel that approaching a vendor might turn out to be a huge success, even if you’re not in the same place as Laura. She, as you’ve realized, didn’t really have much choice: hiring so many in-house CX reps is a mess, and using a bot is just off the table, at least until the technology gets better.

And although this episode made me feel that my company was, as a young startup, a little too strict when it came to in-house approach, I do understand that finding the right vendor is one hell of a challenge, and there are quite a few risks when going down this road. In addition, the team spirit that in-house creates, “the feeling of shared success” – as Roy defined it – is something very valuable for a startup and I feel the advantages of this spirit in my everyday work. So what is right? I couldn’t get a definitive answer, but I was glad to hear Roy say that there’s no need to be too strict about the in-house or outsourcing approach.

I feel in-house allows a lot of things that outsource just don’t, like feeling of shared success. people get to see the bigger picture of things. And even if they need to let go of something or, just like give other people to do something, instead of just them doing it, they can, they will do it if they have a shared feeling of success.

But

If you’re looking for impact, find it wherever you can. If you’re looking for something that will be done well, you know, looking for the best ways to do it. It’s not a question about outsourcing or not outsourcing. It’s not like that’s the goal. That’s not the goal. We just happen to find that working together in the area that we are in is much more effective than having it be distributed.

What I hear is “it’s not about in-house or outsource. It’s about making an impact”. How will you create the biggest impact? Well, now it’s your turn to decide.

Thank you for listening! If you like what you’ve just heard, hit the follow button. You are also more than welcome to our LinkedIn Group, “Startup for Startup”, where you’ll be able to find more content and potential networking. Another place that deserves your attention is our website, startupforstartup.com .

I wanna thank Guy Bin-Nun, our writer and senior producer. Our crew also has the wonderful Adva Shisgal and Darya Wertheim who are our content editors, Lior Krengel leads the initiative, Yoav Ayalon is our opera singer and Gal Roddity is our mix engineer. I’m Dani Lester. More to come, so stay tuned!

[wp_applaud]