Data Engineering Zoomcamp: Free Data Engineering course. Register here!

DataTalks.Club

Starting a Consultancy in the Data Space

Season 13, episode 4 of the DataTalks.Club podcast with Aleksander Kruszelnicki

Did you like this episode? Check other episodes of the podcast, and register for new events.

Transcript

Alexey: This week we'll talk about starting and running a consultancy in the data space. We have a special guest today, Aleksander. Aleksander was a product manager at Delivery Hero. Then he became a co-founder and failed a few startups. Currently, he is running a data analytics agency in Berlin. He was also almost an archaeologist. I’m really curious to hear that story and we will probably hear about that a bit later. Welcome to the show! (1:10)

Aleksander: Hi! I'm very happy to be here. (1:38)

Alexey: The questions for today's interview were prepared by Johanna Bayer. Thanks, Johanna, for your help. (1:41)

Aleksander’s background

Alexey: Before we start with our main topic of starting a data consultancy, let's start with your background. Can you tell us about your career journey so far? And please mention the archaeologist part. (1:47)

Aleksander: Absolutely. I can start with that. Since I was six years old, I think, and up until I was 18 – if you asked me what I'm going to be doing in life, I would tell you, “I'm going to be an archaeologist.” I always loved history and geography. I was collecting dinosaurs and things like that. My entire educational path in high school skewed towards me going to study archaeology, etc. I applied and I even got admitted to two universities. Then that summer, I actually went and worked on an excavation site, to just get a feel of what it means to be an archaeologist. (1:59)

Aleksander: Long story short, when I came back I decided I'm going to stick to the history books and documentaries so I went to study something else, which was political science and international business. That's my educational background. It's pretty much everything and nothing. [chuckles] After that I somehow managed to get into the startup ecommerce ecosystem in Warsaw. I even co-ran a startup on my own before I moved to Berlin, which is also a story. Then I moved to Berlin at the end of 2014. I moved for a job, which was a product management position at Delivery Hero – one of the spin-offs at the time. (1:59)

Aleksander: This is where I met Marco, who is my business partner, or “business better half”. We built pretty cool stuff during our Delivery Hero times together. In 2019, we decided to quit and to do something on our own. We went through a few ideas for a product. That was always kind of the main goal – I wanted to build a product company. I did a bit of consulting in the data space in-between, to also get a feel for what problems there are. Then our main product that we built and tried to market was “data stack as a service,” which is an entire data pipeline from the ingestion to modeling and visualization, including the warehousing. That didn't pan out for various reasons and we decided to… [cross-talk] (1:59)

Alexey: I’m wondering why. I see a lot of startups who work in this area and some of them are doing fine. (4:16)

Aleksander: Yeah, you have Y42, CleanAI – there are plenty of them – MotorData, Veldt. There are quite a few of them. We did a very extensive market validation of that idea. We believe that the market is not there for it – or the market is very, very small. Because essentially what you're selling is a technology – selling technology to tech people is very difficult because they will always think that they can build it faster and better themselves. Even with the tooling, if you look at the options that you have in the tooling landscape today, you can build a data stack in a day or two that's gonna do a decent enough job. (4:25)

Alexey: Can you actually build it or is this just what developers think? (5:14)

Aleksander: I mean, you stitch tools together, right? You take Airbyte, DBT, BigQuery or Snowflake or ClickHouse or whatever, and then you do transformations with DBT and you plug Metabase on top, and you're good to go. The real problem starts when you actually put data into it and you now need to map your business into the data model. And that's something that is very difficult to automate. (5:20)

The difficulty of selling data stack as a service

Alexey: This is probably difficult for a startup that provides a data stack as a service. [Aleksander agrees] Because you only provide the stack and all the business knowledge – all the business transformation – happens after that. (5:46)

Aleksander: Exactly. So if you are selling it to tech people, that's a very difficult sell. Then if you turn around and go to that with the business people – let's say you go to a CMO or a CFO or a CEO and you say, “Hey, look. I have this fantastic piece of technology that you can now use.” And they’re like, “Well, what do I do with it?” “Well, now you need a data analyst to actually run this.” “Okay, now you're selling me something that I need to not only pay for, but now you're telling you to hire someone to do my data for you. Then I’ll just hire the person to do the date for me. So what's the point of my paying?” (6:00)

Aleksander: You would need to build it in a way where you also automate the modeling. Those also exist. I think there is a CLA that automates reporting for marketing, for instance. There are certain tools that automate reporting for finance. Basically, you go directly from the source to a report. That's valuable. Tech on its own is not really valuable. The real work goes into data modeling, because at the end of the day, you just need to map the business into the tables and entities. (6:00)

Alexey: Is this how you ended up doing consulting? (7:09)

Aleksander: Yes. [chuckles] (7:12)

Alexey: Tell us about it. [chuckles] (7:14)

How Aleksander got into consulting

Aleksander: Initially, we had a few other ideas about the product. The first one, we also validated very quickly and it just didn't make any sense. [chuckles] The idea was actually inspired by my inability to pull data from a SQL database back in the days of Delivery Hero and I would always bug Marco to do it for me. We validated that this is not something that would work. Product managers don't have that problem that much. As a community, they have bigger problems to solve first. And then we thought “Okay, so we already quit our jobs. What can we do? Okay, let's go to consulting to actually discover the problems in the data space.” And that's how we started. (7:16)

Aleksander: While doing certain projects, gigs, etc., we discovered that, “Hey, maybe we can do the data stack as a service.” Whenever we would start a project, we very often just needed to build a data stack. It's repeatable work. So that's how it started. We actually found the first customer very, very quickly. From when we started searching for a first customer, it took us two weeks to land one and we had zero product. We had nothing. We basically sold an idea that we're gonna solve that problem. [cross-talk] (7:16)

Alexey: Was this the data stack idea, or the second one? (8:37)

Aleksander: It was the data stack idea. We solved it without having a product and that actually led us to build up the wrong path going forward without getting too many results. So when we decided to stop doing that, we did a really extensive market validation on whether we can sell this, whether the market actually exists. We decided to go back to consulting and fully focus on that from then on. (8:41)

The Mom Test – extracting feedback from people

Alexey: I’m wondering, how did it happen that you managed to sell an idea without having an actual implementation? Was it a part of your user research where you were reaching out to companies to understand what their problems are? And they were like, “Okay, yeah. We have this problem. Solve it for us.” (9:08)

Aleksander: I think this advice is applicable to any kind of business startup – any service that you want to provide or anything like that. Whatever you do, you really need to first see whether the problem that you're trying to solve exists and whether there are enough people to pay for it. So what we did is – I probably need to get royalties for this because I go around talking about this book a lot – the book is called The Mom Test. What the book allows you to do is validate a business idea without having anything – just by talking to people. The trick is that you never tell them what you're doing. (9:26)

Aleksander: Because the moment you tell people what your idea is, they will lie to you because they want to be nice to you and because they know your ego is on the line. It gives very, very practical advice on how to talk to people and how to extract information so that you get very valuable feedback. You do 10-15 conversations like that and that will tell you a lot. Meanwhile, you didn't have to write a single line of code. So we were trying to do Mom tests. We were putting messages in various Slack communities, or different job boards, etc, we were trying to ping people directly. (9:26)

Aleksander: We were just trying to talk to people about the problem of, “Is not having a technology that can process your data – is that a problem that hinders you from actually doing useful stuff with it?” And we offered free consulting for their time. You talk to us and then we will offer you free consulting in return so that the time is also worthwhile for you. We managed to find the person that was our first customer. We talked to them. It was the CTO of a company. (9:26)

Aleksander: They're still our customer, because the product that we built still runs and our customer is still paying for it. At the end of the conversation, he simply asked us “Okay, we told you about our problems. Clearly this is a problem for us. What are you trying to do? What's in it for you?” And we just explained the idea that we had and he basically said, “You know what? Make an offer and you can build that for us. We will be your first customer.” So that's how we sold the first copy of something that does not exist yet. (9:26)

Alexey: So there is a book called The Mom Test. The book describes how to extract information from people – from user interviews – to really understand if the problem you're trying to solve exists and if people want to pay for solving this problem. You tried to do this Mom Test, you reached out to different people, and some of them were quite interested. (12:06)

Aleksander: In general, people will talk to you. People are generally willing to help. If they know that you're not trying to sell them something, they will be very keen. Because people love to talk about themselves – it's just human nature. So if you just know how to nudge and steer the conversation in the right direction, you can really learn a lot. (12:31)

User interviews

Alexey: Can you maybe give us an example? How exactly does it work? Let's say you want to build a startup that sells data stack as a service. How do you approach these user interviews? (12:53)

Aleksander: You need fairly general leading questions at the beginning. For instance, if you're talking to a data analyst – because our initial target audience was data analysts – we knew that data analysts, at the time, very often don't really have enough technical skills to put together a stack like that. They keep extracting, they download the CSV from Salesforce, they have some Google Sheets, they copy and paste – there's a lot of manual work. It's prone to errors, etc. and they try to stitch things together and do analysis on top of that. If they would have a central place where all those sources are in, which is today's standard that's called the “modern data stack,” you can then run your SQL and you can build visualization, and you can help the business to answer business questions. (13:07)

Aleksander: We would ask them, “Hey, what do you do all day? Where does your time go?” You start with that and you listen and tune in for little nuggets of, “Oh, maybe they have a problem here.” And then you dig deeper into that. You never ask questions like, “Would you buy this? Would you pay for this X?” Because if you ask someone very hypothetical questions, it’s like “If you give me this amazing piece of technology – yeah, I'll take it.” Or a question like, “Is this a problem to you?” Like, “Yes, it's a problem.” That's not useful information. People also like to complain. (13:07)

Aleksander: If they tell you it's actually a problem, you still need to gauge how big that problem is. You ask them, “When was the last time it happened? If you don't solve it, what are the consequences? Is it just a mere inconvenience?” Because sometimes it's like, “Okay, whatever. I copy and paste this thing.” But if they tell you, “If I don't solve this, then potentially we're risking losing revenue.” Well, that's a big problem. Then you try to figure out how frequent the problem is, “Is this something that happens every day? Is it something that’s painful, but happens once every six months?” Then maybe it's not big enough. You're trying to extract very, very concrete examples of their workflow, where they suffer, and what the actual pain is. (13:07)

Alexey: How do you record them? I guess that you record the conversation during the interview, right? And then you somehow capture this in a bunch of Google Documents or something like that, right? (15:43)

Aleksander: Yeah. These conversations are usually… we didn't record them. We will take notes. Whenever we could, it would be the two of us doing a call. It's actually very beneficial to do it in pairs. You don't want more than two people. Because if it's just one person that you're interviewing, then there are three people on the call (or in the room) and it might feel a bit overwhelming and stressful, so people won't open up that much, but two people is fine. (15:55)

Aleksander: One person is responsible for asking the questions and nudging in the right direction, and the other person is responsible for taking the notes. The other person also needs to be kind of the guardian of the conversation. It's very easy to start pitching, especially when you already have something. And you never want to pitch them because the moment you reveal what you're doing – again, they're going to start lying to you. Some people are direct. Some people will tell you, “Well your product is stupid. It doesn't make any sense.” (15:55)

Alexey: Most don't, right? (17:10)

Aleksander: But most don't. Exactly. And you can't have 1000 conversations with the hope that you're going to land 10 people that are super honest with you. So you want to avoid pitching as much as possible. Basically, the other person's job is to also make sure that whoever is asking questions doesn't get too excited and jump ahead with pitching too early. (17:12)

Alexey: Your background is in product management, archaeology aside. [Aleksander chuckles] What is Marco’s background? (17:38)

Aleksander: Marco is a software engineer. (17:47)

Alexey: Software engineer. Okay. I guess it was the two of you – you and Marco – who were doing the interviews. So how did you split these responsibilities? Who was taking notes? Who was asking questions? (17:48)

Aleksander: It depends. We could switch. (17:59)

Why Aleksander’s data stack as a service startup was not viable

Alexey: Oh, so you could do both. [Aleksander agrees] Interesting. So you did these 10-15 conversations following the Mom Test and then at the end found out that the problem is not big enough? (18:01)

Aleksander: Yeah. First off, the mistake that we made was that we sold it without having it – which is an ideal scenario. If someone is taking out their credit card and telling you, “Yeah, I'm gonna buy this without you having anything at all.” That's a great signal. You just need to have enough people that will do that. [chuckles] We got excited, we went and built it, we launched it, we did everything, etc. We helped with setting it up, with the models, and with setting up the Metabase. They already had Metabase, so we transferred some of the reports onto the modeling layer so that they were faster – so that they actually run. I was like, “Well, if we can find a customer in two weeks without having it, we shouldn't just go out and spam everyone. We’re sitting on a goldmine, right?” (18:18)

Alexey: Was this the case? (19:21)

Aleksander: No. It wasn't. We spent a good few months trying to acquire the next customers and it never happened. That's when we were like, “Hang on. Maybe we were trying to run before we could walk. Let's go back to validate.” And we went back to validate and we managed to line up I think 15 people that could be potential customers. Then after talking to them, doing the proper monitors, etc it became kind of obvious that it's just not there. I mean, this product makes sense. It makes sense and there are companies that will use it. It's just that I don't think there are enough people – enough companies where it makes sense – or that the lifecycle is long enough. (19:22)

Aleksander: You sacrifice a certain degree of flexibility with products like that. It's a black box. You don't know what it is. You can either build it in an opinionated way or you build it in an open way. But in an open way, it's either open source, or there are different pieces of it, like ingestion, transformation, etc. Also, making compromises is very difficult. You can’t have it be inflexible, yet allow for certain changes. That's just very technically challenging to actually build. So you need to pick. And if you're picking it as “This is what it is. This is a black box. You don't really ever know how it works.” People don't like that. (19:22)

Aleksander: We talked to some of the customers of our competitors and there is something to be said. If your team is a little bit more junior, then this provides certain guide rails and they cannot mess up too much. I mean, they can still mess up models, but you can always mess up models. So having that kind of rigid structure can basically make up for certain mistakes that less experienced people make while working on the pipelines. (19:22)

Alexey: This client of yours that you landed while conducting this Mom Test – they were interested in the product that you wanted to build, but then you kind of ended up consulting them, right? You said, “Let's build this thing for you.” (21:39)

Aleksander: Exactly. Because what they really wanted, and what the real benefit was – it was the free consulting that they got out of it when we were actually building it. We built their models. We went and built their modeling layer. It was in our tool, obviously, but we wrote the SQL. And that's the real value. We went and we understood their business and translated it into SQL models. That was the real value – not the shell that would run it. It took us a while to realize that, actually, because we were excited that someone is actually willing to pay for the product that we bought. [chuckles] That can lead you down the wrong path. (21:56)

How Aleksander decided to switch to consulting

Alexey: How did you realize that building these models is what you want to do and that you and Marco want to start this consultancy company? (22:42)

Aleksander: It's where the value is. When you're working with the data, to us, it’s bridging the gap between the business and whatever is hidden in the data that you collect. First and foremost, you start with that. Someone has a business question they need to make a decision, you start with that. You spent time understanding what it is. It's a lot of talking to your stakeholders. There's a lot of product work there. I always say that – there's so much overlap between the product managers’ job and data analysts’ job, because data teams don't usually have product managers. Some bigger ones and bigger companies, of course, do. But most of the teams don't have that luxury. So someone needs to go into that work. (22:50)

Aleksander: At the end of the day, data teams –it's an internal product. If you're talking about business intelligence, it's an internal product. You are serving your internal customers and they work in other teams. This is where the value is, and this is what's difficult. There are a lot of toys now, we have ChatGPT now and there's a lot of talk about data science, etc. A lot of companies can do without it. A lot of companies don't have the data and they need help with basic things. The basic things take them 80% of the way. And once you get that 80% of the way, then you can start thinking “Okay, how do I optimize the 2%, the next 3%, the next…?” (22:50)

Aleksander: This is where machine learning and this is where data science really comes into play, and really can be valuable. But at the beginning… and we're talking here about medium-sized companies, we're talking about startups, post-seed, pre-series A, series A – that kind of level, unless your value proposition requires AI and requires machine learning. That's a different story. Then you really need to invest from the get-go, because otherwise you don't have a value proposition. But when you're talking about analyzing the business, understanding how the business is doing, and predicting how the business can be doing, there's a lot of very simple and pragmatic and very often boring things that you can do that do not require a lot of effort. But to dig this out, you really need to understand what is needed and that's the difficult part. (22:50)

Alexey: I want to ask you about these simple, pragmatic, and boring things. But maybe before that, I’m curious. Why is it called consulting? Do you just go there, ask questions, or are you hands-on? (25:45)

Aleksander: No, we are hands-on. We are implementing things. Consulting can have a very, very negative connotation. Let's be upfront about it. You pay consultants, they come in, they tell you what to do, and they go away. They give you their invoice and wash their hands of the consequences or the results. No, we implement things. We basically go in and we will try to figure out what's needed, and then we will implement those things. We take full accountability for if it works or not, obviously. I mean, in a nutshell. [chuckles] (26:03)

Alexey: So why is it called consulting? To me, consulting is actually – a company has a question and then they need to consult somebody on solving the question. They consult, they get input, and then they go implement it. I guess this is like the classical definition. (26:44)

Aleksander: Well, this is a good question. Why is it called consulting? That's what it means – you consult someone if you have a problem and they basically get you unstuck. But we're also talking about implementing things. I don't know. (26:59)

Alexey: This is called consulting too, right? And this is the kind of consulting you do. (27:16)

Aleksander: Yeah, it's just a subgroup of consulting? A type of consulting? Probably. But yeah. Very often, I laugh about the fact that you're being paid to comment, advise, and then implement what people need, but very often, you end up just implementing what people want. There is a limited amount of influence you can actually have and exerting influence, sometimes it's just difficult. It really depends on the client. (27:19)

Finding clients to consult

Alexey: First, let me ask you about how you find these clients. You said that your clients are usually startups – pre-seed startups or series A startups – that do not necessarily have this data infrastructure. In order to do cool things in the future, they need to get the basics first. So how do you find these companies? Where do you find them? Do you open TechCrunch or something like that? (27:59)

Aleksander: No. [chuckles] It's a little bit different. Consulting or freelancing or contractual-based work – because, at the end of the day, we're external partners, external providers, or external contractors – it's very much a network-based business. But you cannot just wait until your network gives gigs. You need to help it a little bit. Getting the first projects is usually the most difficult part. (28:33)

Alexey: For you, it wasn't that difficult, right? (29:14)

Aleksander: When we switched to consulting, it's a little bit different. We would tell people, “Hey, we’re doing consulting now.” You reach out to your network first – that's what you do. This is the first step. Basically, you let people who you think can introduce you to potential clients or could have worked for you. You build up a network and people may go work for different companies – maybe they need help as well. So you do that first and let people know that you are looking. You need to think about, “What is my positioning? Who is actually my target customer?” You need to explain to others what you're doing, even if it's just consulting, it can have many shapes and forms. So you need to think about that and reach out to them. That's step number one. But again… [cross-talk] (29:18)

Figuring out how to position your services

Alexey: How would you actually do this? It sounds easy, right? You're just like, “Okay. Who is my target audience? How do we position ourselves?” It seems simple. But when you actually start, it's not easy, right? How do you actually do this? Mom Test? (30:17)

Aleksander: We struggle. Basically, we experiment. We see what lands – what strikes a chord on the other side. You always start with, “Okay. Who do I want to work with?” Even now, we are experimenting with how we position ourselves. For instance, back when we were starting, we thought, “We’re the guys who will come in and build your first data warehouse, or first reports, models, etc.” That's what we were doing. We knew which moment in the lifecycle of a company where this becomes helpful and actually makes sense. (30:37)

Aleksander: Because you shouldn't build your data warehouse too fast – there is a point in time where that makes sense. At the beginning of the company, it just simply doesn't. So we knew which kind of companies we should go for. These were around series A startups. If we're doing things, this is where you're getting your product market fit, so this is where you're going to start scaling. And that's something that is really useful – when you more or less already know what you're doing and you just need to do more of it. You're getting out of the experimentation phase. This is where… [cross-talk] (30:37)

Geographical limitations

Alexey: Did you have any geographical restrictions? Did you focus on companies in Berlin or you just targeted everyone? (31:59)

Aleksander: Europe – as we're in more or less the same time zone. Working with companies in California would have been difficult. Also, we don't have a footprint there. We don't have a network. Our competition is always local. You can go global, but when you want to grow to a certain size it probably makes sense to expand. We would talk to VCs – we know a few VCs, etc. – “Hey, does anyone from your portfolio companies need help?” You think about which people can help you earn the clients. I just want to go back to positioning because that was your question. You need to think about, “Who is your customer?” And “What is the value proposition? What are you offering them?” Even in the data field, it can be so many things. (32:05)

Aleksander: Now we're trying to say, “We are on demand data analysts. We work directly with the business. We can help you optimize your revenue. We can help you optimize your marketing spend. Maybe there is some kind of risk that you have in the company and you want to mute that risk. Maybe you want to know who your customers are. Maybe you want to do some kind of segmentation.” We start with a business opportunity or a business problem. Yes, we can build the infrastructure. We have full stack capabilities, between the two of us. It's a blessing and a curse, because we can literally do anything. We have product skills. The both of us have data analytics skills. And then Marco has the engineering skills. So we can really do a lot. (32:05)

Aleksander: But you can't just tell people, “Oh, I'm doing everything!” Then you’re probably doing nothing. You need to think about what you want to offer. You can say, “You know what? I'm a data scientist or I'm a data engineer. I'm gonna partner up with someone and we're just gonna hit up the companies that already have big teams. They just need hands-on support on certain projects.” You can do that. That's the positioning. It just comes down to what you can do, what you can do well, and who you can help. It's difficult. You have to choose the right words. (32:05)

Alexey: So first, you think about “What do I want to do? (If I'm a data scientist) I want to help companies who already have a data science team, but maybe they're short on people. I can go there and help with their existing projects. Or I can target companies who do not have any data infrastructure, but they want to do some data stuff in the future.” First, you need to define who your target customer is, and then once you define that, you need to come to them and tell them why they should hire you. In your case, they probably have some business problems or things they want to solve like, “Who are our customers?” (34:42)

Alexey: Maybe they need segmentation or optimize their spending on marketing. I guess, in this case, you need to know the domain in which they work a little bit? Then you can speak the same language with them. You can already suggest some things. This is your messaging. This is your target customer. You already have some problems in mind that you can solve for them. This is what you pitch. If they like you, they hire you – if they don't, you keep iterating on your messaging, right? (34:42)

Aleksander: Well, yes. Essentially, you can keep iterating on your messaging. You can also keep iterating on your positioning. As I said, “What is your value proposition?” (36:03)

Figuring out your target audience

Alexey: What is that? In these two things – the first is the target customer, and then there’s the thing that you can solve for them. Is positioning related to the first one or the second or both? (36:20)

Aleksander: It's related to both. It includes who your customer is and what your offering is. From that, you can figure out “Okay, where do these people hang out? Where do I find them? Who do I know that can introduce me to those people? How do I want to structure my messaging around that?” That's part of positioning as well. (36:33)

Alexey: Can you give an example of this messaging? Maybe this is what you’re suggesting to your customers right now? Maybe you can give an example of that? (37:03)

Aleksander: Yeah, exactly. For instance let's say we're on demand data analysts. That's still very broad. Data analysts can do a lot of things. So what can you help me with specifically? If we're talking to a CFO, you can say, “Look, we can help you optimize your revenue. These are tight times. Especially if we're talking about VC-backed startups, investors are looking at their PMLs with more scrutiny right now. Let us help you find opportunities to either reduce your spend or to increase your revenue. If you need to optimize how much you're spending on marketing, let us help you with that. We can figure out your tracking. We can try to optimize where your marketing spend should be going – how you calculate your return on investment, etc. We can do something around attribution.” It really depends on who you're talking to and the problems that they have. (37:13)

Alexey: I guess you need to do a bit of homework, right? Before you speak with a customer, you want to understand, “Okay, they're in this particular market segment. They probably have this, this, and this problem.” Then you go to a meeting with them and you say, “This is what we can help with.” You start talking with them, and while talking you realize that maybe the problem is not what you wanted to suggest, but something related. While talking, you figure this out. (38:22)

Aleksander: Yeah. Going back to how you acquire the customer. That is relevant to, again, the positioning, etc. Right now, we're testing a few things. Last year was fantastic. Last year, we were rejecting projects, because we just did not have capacity. And last year was still a year where there was a lot of VC money and people were still hiring. At the end of the year, we saw a lot of layoffs in various companies. That also includes people in the data industry. Because of that, there was an influx of people into the market. Yes, there were layoffs, but there were still a lot of companies that were hiring. You could still find a job. The competition on the market became bigger when you were a job seeker, but there were still companies that were hiring. Those companies that have the budgets and where problems need solving, six months ago they could not hire. (38:47)

Aleksander: This is where an external contractor becomes very attractive because you can have that person tomorrow. Trying to poach and things like that – there was just not enough supply for the demand. Now the demand is much smaller and supply has increased. Therefore, the companies that were cutting budgets, maybe taking contractors doesn't make much sense. However, it actually might, in some cases. People still have certain needs to be fulfilled – they can just hire easily now, and probably cheaper because the salaries are also inflated. Now we're probably going to see that the market is shifting towards the employer market. (38:47)

The importance of networking and marketing

Alexey: Is there still room for consulting right now in this market? (40:50)

Aleksander: I think it's tougher, yes. One of the cons of doing this is that there is uncertainty. Yesterday, today, you have projects – tomorrow you might not have them. Essentially, it is up to you to find work. But yeah, I believe that it's possible. We do have a little bit of work. We have been talking to people. We have leads. People still need support. But this is where you position yourself and things like that become crucial. You rely on certain uncertain distribution channels, let’s call them that, and it's not like they are super repeatable. You can't just say, “I'm going to spend money here and I know I'm going to get clients.” It doesn't work like that. I mean, you can do it with outbound sales, but if you're small, that might not necessarily make a lot of sense. (40:55)

Aleksander: You network – you message people, you talk to them, you wait. We do content. LinkedIn works, actually. It's very interesting. If you put yourself out there with a very constant messaging that relates with your target audience, we've seen that people actually reach out to you to solve their problems. That's why it's important to play around with the message and then build the content that you produce around that. [cross-talk] (40:55)

Alexey: Katya, who put us in touch says exactly the same thing. By the way, if you're listening to us right now – Hi, thanks for putting us together. I think she mentioned what you’re saying as well. (42:24)

Aleksander: We know this from her. We collaborate on a few things. That's why messaging is important. Content, in general, works. We also had people getting in touch with us in the past because someone saw a blog post that we wrote and then someone asked that person if they knew anyone and that person replied “I actually read a blog post. You know what? Go talk to these guys. Maybe they can help you.” It’s just that building this up takes time. But in the long run, it pays off. You do those things – you do all of them – and it all compounds. That's how to build a stable engine to be able to talk to people. (42:35)

Alexey: So you need a network and with the network, you get referrals. Happy clients refer other clients. But you also need other sources of leads. One of these sources is content. Content could be, as you mentioned, LinkedIn posts. You post messages targeted at your ideal customers – your target customers. Then you also write blog posts that are also targeted at these customers. By the way, do you make these posts yourself? Or do you have some sort of content manager? How do you do this? (43:21)

Aleksander: No, at the moment we make them ourselves. When we started writing blog posts, we worked with a freelance editor who would help us. Writing is difficult – it's a skill on its own. (44:00)

Alexey: I guess for you, as a product manager, it's easier because… [Aleksander disagrees] No? [Aleksander laughs] For software engineers it's probably harder than PMs. (44:17)

Aleksander: It's a skill. If you're not doing it on a regular basis it's… it has rules that you need to follow. There's a lot of things to consider. So we worked with someone that would help us structure the articles, read through them, proofread them, point out mistakes etc. That helps because you learn from that person as well. You get tips. You talk to that person, who says, “Hey, maybe you should structure that in a different way.” The person that you work with doesn't need to know anything about what you're doing. You need to explain. Today, we still write our own content. I like writing my own content because it allows me to also express who I am. It's mine. So it's kind of important to me that I write that. (44:25)

Pricing your services

Alexey: There is a question that has been there for quite some time. I think now it's time we covered that. The question is, “How did you decide on pricing for your first client?” (45:19)

Aleksander: Um… [wets finger and sticks in the air] I'm gonna go with that. (45:33)

Alexey: For those who listen to this without the video, what did you just do? (45:39)

Aleksander: It was a more or less random number. Sorry, it wasn't random. It was a number, it was not random. We looked at what competitors are charging. The way you need to think about this is – you want to base your pricing based on value. You don't base a pricing of something based on what it costs you to produce it. That's the first thing. If you're trying to design a product or a service, that’s the first thing you need to understand – the service does not cost the client or customer the price that’s based on the “manufacturing costs”. It's based on the value that whatever you're doing is providing. Our thinking was more or less along those lines. (45:44)

Aleksander: We tried to enable data analysts. At the time, hiring data engineers was extremely difficult because these were the times that data engineering was so hot that everyone wanted to hire one. I even did benchmarking and I did some LinkedIn searches to just quantify how many jobs are out there versus how many people there are with that title. The gap was massive. In Europe, it was 100% and in the States it was like 200%. In the US, even if everyone quit, or was fired, or was poached, and switched jobs, the gap would still exist, which is ridiculous. So, “A data engineer costs X, I'm saving you the full-time hire.” The service, obviously, needs to be cheaper than a full-time hire. That's the rule of thumb. (45:44)

Alexey: “Cheaper” in what sense? I’m guessing that you, as a consultant, need to charge a bit more than, let's say, a month of a data engineer’s work. (47:33)

Aleksander: Oh, sorry. I was referring to selling a product. You ballpark, “Okay, maybe it should be this.” And then you say, “Okay, what is the competition charging?” At the beginning, getting the pricing rate is not that crucial. At the beginning it is more important to see if whatever I'm selling sells and people are willing to pay for it, then you will have to change your pricing on the way, once you get more data points on what works and what doesn't – when you learn what the value is that you provide. For consulting, the way we look at this is, although we charge daily rates, it's also not that someone is just paying us for our time. Someone is paying us for the fact that we've seen things multiple times and we will be able to tell people and navigate a situation that they have never seen before but we did. (47:42)

Aleksander: That's why the pricing is based on that. The other thing is, when you're hiring externally – we can be fired at will. Tomorrow, you can tell us “We don't want to work together.” And that's perfectly fine. That's where that premium goes as well. For us, we compensate for that. We're not full time employees. Especially in Germany, when you pass the probation period, it’s very difficult to let people go. So you need to include that. That's why you pay the premium on that. (47:42)

Alexey: Is there any rule of thumb for that? Let's say, “Take the salary of a senior data analyst and multiply it by two,” or something like that? (49:18)

Aleksander: No. I mean, you try to figure out what the market price is – what the hourly rate is. (49:28)

Alexey: How do you do this? How do you figure it out? Do you look at how much other consultants charge? (49:35)

Aleksander: Yeah. We talk to each other. The community is there and we talk to each other. With the ones that we know, at least we kind of know what the ballpark pricing is and what everyone charges. And that's okay. When we first started, we had one of the first clients – it was a person that we knew. He was also giving us advice on startups and how to build a company, etc. He actually had experience consulting as well. We quoted him and he goes, “So how much do you charge?” And we said, “Okay, we want this.” He looked at us and said, “You know what? I'm very happy to take that price. Next time you do this, double it.” (49:42)

Alexey: [chuckles] Did you end up doubling it for that person? (50:33)

Aleksander: No. For the next clients, yeah. The other thing that is really good to think about is – figure out what your maximum is. What is your starting rate? What is the ideal that you want to charge? Obviously, it cannot be ridiculous. You can’t say, “I'm gonna charge 1000 per hour,” that's just not gonna fly. It's impossible. But you need to figure out. [cross-talk] (50:40)

Alexey: You can’t? Oh. Too bad. [chuckles] (51:03)

Aleksander: I mean, maybe it depends on what you do. [chuckles] (51:06)

Alexey: But not in the data space. Right? (51:10)

Aleksander: Not in the data space, no. [cross-talk] If you figure out how you can sell a one hour of your work for a 1000, do let me know. (51:12)

Alexey: I guess, usually, it's like 1000 for a day, right? Or something like that? (51:21)

Aleksander: Yeah, that would be the ballpark. And that depends on what kind of work you're doing and how long the collaboration with the client is going to last. So this is your starting rate and then you also need to think about the absolute bare minimum that you are willing to take the job for. This is useful because when someone asks you about the price, and even if it's a friendly company owner – someone that comes from the network or from close friends or whatever – and they ask you, “Please make me a good deal.” Never, ever start with a good deal. Because no matter what you quote, people try to bring it down. We made that mistake a few times. (51:26)

Alexey: Especially good friends, right? (52:14)

Aleksander: [chuckles] We've made that mistake a few times, so right now we're always starting with “This is our starting point.” And then no matter what it is, people always try to bring it down. You start with the high one and then you basically end up more or less in what is acceptable. (52:15)

The pitfalls of daily and hourly pricing and how to balance incentives

Alexey: With the daily price – I have no experience in consulting, just something that's on my mind – you have this incentive to work for more days. So instead of thinking about, “What is the most effective way to solve the problem and have a solution?” You start thinking about these days like, “Okay, I need six months for that. This is how much it will cost.” Right? You kind of create the incentive (that maybe you don't even realize because it's our internal bias) to work more. [cross-talk] (52:38)

Aleksander: Yes. This is an excellent question. You’re right – there is an incentive to work more days. I'm being very honest, we really try not to do it. We try to be really honest and say, “This will take X days. Maybe it will take more, maybe it will take less.” Our incentive is, at the end of day we want to help. There's also a very big incentive for you to do a good job and for the clients to be happy. Because if the client is happy… [cross-talk] (53:15)

Alexey: You get referrals. (53:54)

Aleksander: Exactly. Because if the client is not happy, they're not gonna refer. This is a small world. Everyone knows each other, especially in the startup ecosystem in Germany, and even in Europe, in general. People talk. The fact that you didn't do a very good job, people are going to hear about it and you don't want that. You want them to be happy. You want them to refer you. Because no one goes into Google search (or these days, Bing, since Bing is a thing again) and says “Okay, I need data consultants. Who are they?” No, you go to someone that you know and ask, “Hey, do you know them?” That's how this search happens. So you want happy clients. (53:55)

Aleksander: If your incentive is to ride as long as possible and ride on the client’s inexperience or the fact that they don't know how certain things work or they don't know the tech or they don't know how long it's gonna take – it's gonna bite you at some point. So there's incentive to avoid that and there's incentive to be honest and actually helpful to your customers. Then there is another thing – you can turn around and say, “Okay, let's charge for a project.” The problem with that is, in general, it's very impossible to know what a deliverable should cost because you don't know how much it's going to take before you actually start doing it. (53:55)

Alexey: You need a few iterations, right? You cannot just come up with a project proposal where you deliver and you shake hands you part ways. [Aleksander agrees] You would always need to make a few iterations. (55:33)

Aleksander: If there is manual work involved – and you can automate certain things, that's obviously possible. You can build tools for yourself and make your next project easier if it's the same type of a project. The main effort that we put into our projects really goes into data modeling and that mapping of the business into SQL tables and business entities, etc. so that you can actually make this useful for the rest of the company. You don't know how much time it's going to take. You can ballpark things, of course. If you have one source, it's probably going to be easier. You just need to figure out that one source, right? (55:45)

Aleksander: If it's your application database, “Okay, I need to talk to developers. If you're an e-commerce business or delivery company, when I look at it, I will know what I'm looking at because it's not the first time I'm looking at it.” But there'll still be bugs. There'll still be weird things that I will need to go and figure out. I still need to talk to stakeholders, and some of them are not super available, because they're also busy people, so you might be rescheduling things. You need to get a feel for how the stakeholders like to be communicated with. You need to get a feel “Okay, how do I get the information that I need from those people? There's always that kind of uncertainty, because you never know what you’ll discover. (55:45)

Aleksander: It's impossible to project how much something should cost based on the effort that is going to go into it because you just don't know. Then at some point, if you do these projects, there's always one side that is unhappy – either it took less time because it's possible. It's very variable that you discover “Ah, it’s super easy. I will solve it in a week.” And we're charging something like for a month of work. Then the client will be unhappy because now the client thinks they're being overcharged. Or you spend significantly more time on it and then it's just unprofitable for you. (55:45)

Is Germany a good place to found a company?

Alexey: We don't have a lot of time so maybe I'll ask a very quick question with a yes or no answer. Is Germany a good country to register a consultancy company? (57:58)

Aleksander: [laughs] Yes and no. As a good consultant would say, the answer is yes and no. If you want to primarily target German small/medium-sized companies or German enterprises, then yes, Germany is a good place to found a company. Because those clients will treat you seriously or actually will consider you at all, if you have a company in Germany. Ideally, that company is GmbH, which is a complex form of incorporation, especially in Germany. (58:10)

Aleksander: If you don't care where your customers are, like in Europe, or your customers are primarily startups, then you can consider other countries. You can do it in the Netherlands, which is very easy. You can do most of it online. It's very quick. Everything is in English. You can consider Estonia because then it's just simplicity. Doing it in Germany has a certain amount of bureaucracy that comes with it, which we learned the hard way. (58:10)

Alexey: Where are you registered? (59:21)

Aleksander: In Germany? (59:24)

Alexey: In Germany. [chuckles] Okay. (59:28)

Aleksander’s book recommendations

Alexey: Last question. Are there any books or other resources that you can recommend to our listeners? (59:29)

Aleksander: Other than The Mom Test? Yes. There is one that is very interesting that is called Think Like a Rocket Scientist. And it's written by an actual rocket scientist. The guy was working for NASA and he was working on the operational team that was sending things to Mars. Basically it deconstructs a way of how you approach projects and how you should be thinking about projects. There's a lot of overlap with how you build software, “You should test. You should test. etc.” But it basically breaks down how you should start with your problem or the big idea and then kind of reverse engineer and formulate a plan from the top down, rather than just trying things randomly. It's written in a very easy way and there are really cool examples and real stories from space missions and how NASA screws things up. It's an easy light read but there’s actually a lot of good advice. If you don't have extensive experience in the product management field – because I do those things on autopilot now – but it actually teaches you the kind of thinking that is needed. Starting a company is a project. How would you go about that? There is a lot of good advice there. (59:35)

Alexey: Thank you. That's all we have time for today. Thanks a lot, Aleksander, for joining us today and for sharing your knowledge, expertise, and everything you learned from your experience of starting a consultancy company. Thanks, everyone for joining us today, for listening in and for asking questions. Have a great weekend, everyone. (1:01:26)

Subscribe to our weekly newsletter and join our Slack.
We'll keep you informed about our events, articles, courses, and everything else happening in the Club.


DataTalks.Club. Hosted on GitHub Pages. We use cookies.