Data has been the talk of the town for product managers over the past few years. To provide some insights on data as a product, how to manage data products, and conclude whether it’s pronounced ‘Dayta’ or ‘Daata’, we spoke to data expert, Omar Sallam, Senior Product Manager at Booking.com.
Featured Links: Follow Omar on LinkedIn and Twitter | Omar’s Product School ‘How to manage data-as-a-product’ webinar | ProductTank Cairo
We’d love to hear your feedback on how we can improve The Product Experience podcast. Let us know your thoughts by filling out this quick form or alternatively, share your thoughts on this episode in the comments section below.
Randy Silver (00:00):
Lily, today we’re going to talk about one of the most controversial subjects in product management. Are you ready?
Lily Smith (00:06):
No, I don’t want to do that today, Randy. I can’t talk about dates on roadmaps anymore.
Randy Silver (00:14):
Okay, we won’t do that. But the subject I had in mind was actually a lot more contentious than that. We’re going to talk about data as a product or is that data as a product.
Lily Smith (00:27):
Oh, okay. That one is more contentious, but I can answer that. It’s definitely data.
Randy Silver (00:35):
Okay. Yeah, well, you’re correct and that was a little bit easier than expected, but we did bring someone in to talk to us about this subject, not about pronunciation, but about actually managing data, not data, data as a product and the difference between data as a product and data products, which itself is weird. So, our actual expert is Omar Sallam. He’s a senior product manager at Booking.com.
Lily Smith (01:02):
So, before Randy says data again, let’s get cracking with the interview. The product experience is brought to you by Mind the Product. Every week on the podcast we talk to the best product people from around the globe. Visit mindtheproduct.com to catch up on past episodes and discover more.
Randy Silver (01:21):
Browse for free or become a Mind the Product member to unlock premium content discounts to our conferences around the world and training opportunities. Mind the Product also offers free ProductTank meetups in more than 200 cities and there’s probably one near you. Omar, thank you so much for joining us today it’s great to have you here.
Omar Sallam (01:43):
Yeah, it’s great to be here. Thank you so much for having me.
Randy Silver (01:47):
So, for anyone who doesn’t recognise your name already, just give us a quick introduction. What are you up to these days and how did you get into product in the first place?
Omar Sallam (01:57):
Well, my name is Omar Sallam. I’m a senior product manager at Booking.com. I’ve been with Booking for a year and a couple of months now, and I’ve been in product for about six, nearly seven years. I started out my career about 13 years ago as a system analyst back in my country in Egypt. And at that time, product management wasn’t still a thing, especially in the Middle East, but that was the normal position for what evolved after that to be product management. So, I did that for a couple of years, went into project management, and then I finally found product management, which was the best fit for what I was always hoping for as a career. I worked with a couple of very impressive startups in Egypt and then started getting into larger scale companies. Most recently before Booking.com, I worked with Jumia and then Delivery Hero in Middle East Talabat, and then I joined the Booking at the beginning of last year.
Randy Silver (03:05):
Fantastic. Let’s put it kindly, I think data obsessed, it would be a good way of putting it. So, we’re going to talk about data as a product today, but what kind of data and why is it so important? Why do you focus on that?
Omar Sallam (03:24):
Well, I’m focused on that idea because my current role in Booking.com, which is working with the location intelligence team working on geodata, which is a very specific kind of data, made me more interested to know what are the best ways to manage data as a product and to understand this concept a little bit more. And by the way, this was my first role to as a data product manager. I feel that I was really lucky throughout my product experience that I worked on always on different scopes and different kinds of products. And I like to think of myself as a generalist product manager, and I think this is something that actually has made me understand better the concepts of managing data as a product, because it’s all about applying some fundamentals and frameworks and general ideas to be flexible, to be applied in any way.
Just like you can apply the frameworks and ideas of product management on any kind of scope or problem spaces or product domains. It goes the same for considering data as a product. So, as I told you when I started my current role on a very specific type of data, you have to start thinking about how to consider this data set as a product itself. What are the characteristics of it, how it can evolve, how can you drive impact? Because when you’re working on data that is purely infrastructural and you’re trying to serve this data to different teams that are using it for several of their use cases and different experiences for the customers and so on, you always try to proxy your impact through these teams that you are servings and these products that you’re empowering. So, it’s not easy and it’s not straightforward to understand what kind of impact you would be driving to the end customer.
Lily Smith (05:33):
Omar, just taking it back to the beginning there, like you said that this was your first role as a data product manager.
Omar Sallam (05:40):
Lily Smith (05:40):
I’m just curious to know what drew you into this role. What kind of excited you about being specifically a data product manager? And also, you’ve been in the role for over a year like you said, so are you seeing that it’s different to normal or other types of product management? Is there any sort of particular skills or anything that you need to be specifically a data product manager?
Omar Sallam (06:07):
Okay, so what drew me into that, as I told you throughout my experience, I’ve always worked with different kinds of products, different domains, different businesses, and I always appreciated and got intrigued by a new challenge. So, I really liked and got interested into the description of the role itself. And since I was a kid even, I was always excited about geography and maps and everything, so that’s why I was really interested to get into this. It was a totally new space for me, but it sounded like a new challenge that I wanted to take that’s the main thing. So, when I started, the main things that you need to understand as a data product manager is it’s not something specific. So, you don’t have to have a lot of technical skills in data for example, it would be definitely a plus. But the most important thing is actually understanding the concept of how to treat data as a product.
And that’s why I started digging more into this because as I always think, you have to proxy your impact to the end customers. You are serving a lot of teams internally inside an organisation and you are trying to see how to evolve the data that you are managing, the quality of the data, the outputs and processing of this data to serve various use cases across the business to deliver impact through other teams and other experiences that you don’t have control of. So, you have to have a very wide holistic view and holistic style of managing the product itself in order to think about how to strategize about it, how to build a roadmap and how to prioritise the things to work on to be able to proxy that impact.
So, it’s not really about technical skills and having a technical background and data management, that would be a huge plus. But the most important thing is how to be able to understand the impact that you can proxy. And that’s something that I think I was lucky enough to be exposed to different kind of experiences and different kind of products and scopes because the more you are working with different use cases and different scopes and domains, the more flexible and the more holistic your mindset gets. And that would definitely help anyone that works as a data product manager.
Lily Smith (08:37):
So, what are the biggest challenges when you’re working as a data product manager?
Omar Sallam (08:43):
The biggest challenge is the nature of data itself. The concepts of managing data as a product are relatively new, just started to evolve in the last three, four years, five years. And historically, in most companies and most organisations, however, the size is the best challenge with data is mostly it’s extremely fragmented. It’s really distributed between different systems and different teams, whether it’s in-house or you’re using third party tools or integrating with them, data is scattered everywhere. Or if it’s a little bit structured, again between different teams, between different business units, you will definitely find different structures and different standards. And as the company size grows, those structures tend to be even more different and more and more various.
And of course, there is always the different kinds of compliance policies and regulations that you have to follow. And it’s not always a hundred percent applied across an organisation, especially also if the size of an organisation is large. So, there will always be some business units or some departments that are more advanced into applying such policies or compliance programmes and others are not, and structures are not compatible with each other and so on. These are the main challenges and the biggest challenges because if you’re trying to work with data and serve it horizontally and try to maximise the impact you need to make all of this work together somehow.
Lily Smith (10:26):
It’s interesting because when you think about data as something, it touches every single person in the business in some way or another. And so, it’s almost like being the product manager for communication or I don’t know, motivation or something quite sort of intangible if you like. So, you’ve mentioned a few of the challenges there. What are some of the ways in which you’ve kind of overcome those challenges and tried to understand, I guess, the landscape within the business that you’re working in so that you can build up a picture and know where the priorities are?
Omar Sallam (11:03):
So, as I told you, being a data product manager is about serving a company horizontally you’re serving all the teams, all the business units, all kinds of experiences. So, one of the main things and the most important things is maximising your communication efforts and stakeholder management efforts. You have to speak to everyone. You have to be involved with all the strategic planning and strategic thinking and discovery efforts for the different teams that you are serving and also start getting more into the activities for the teams that you’re still not serving yet. Aiming to discover opportunities of collaboration. This is something that’s not easily done and it’s really hectic and complicated both on your side as a data product manager and also for your stakeholders because it has to go both ways. They have to support you, they have to explain different use cases for you, explain the whole thing as if you are one of their team, so you’d be able to identify these opportunities.
So, communication is key always being a hundred percent focused with depending on the size of the company, it might be tens of different teams and being totally focused with them and identifying all of these different opportunities and structures and trying to make all of this work together. This is the most challenging thing because of the other side, you are trying… As I said, to proxy or impact. So, you’re trying to put the different teams and how they are planning to drive impact on their metrics and KPIs into groups and trying to come up with four or five different funnels, grouping a couple of teams together to be able to prioritise your own things and focus on these kinds of impacts or metrics that are grouping different teams together.
So, to sum up communication as the most important thing on the other side, the deeply technical aspect of data can be balanced between you and your technical team. So, as a product manager, just focus as much as you can to create these channels and funnels of driving impact to be able to strategize about the product itself and how it evolves and how it grows, and to be able to serve more and give space and empower your technical team to also take their own decisions to support you enough to be able to focus on the product aspect of things.
Randy Silver (13:47):
Are you ready? For Mind the Product San Francisco conference happening in June.
Lily Smith (13:53):
If you’ve been before, you’re probably feeling a bit like me, desperate for your MTPCon fix. And if you are new to it, this is the product conference not to miss. If you’re a product person looking to advance your career, expand your network, get inspired and bring the best products to market, then this is for you.
Randy Silver (14:13):
So, what can you expect? Well, MTPCon is known for their epic lineup of speakers, renowned product leaders with invaluable insights and tactics to share. They cover a range of exciting topics that will challenge and inspire you to step up as a product manager always with something tangible to take away into your own product practise.
Lily Smith (14:35):
And don’t let location hold you back. Even if you can’t join in person in San Francisco, you can still be part of the action with their convenient digital only option.
Randy Silver (14:47):
This event is a must attend for anyone seeking to elevate their product management game. Find out more and book your ticket at mindtheproduct.com/san-francisco. Omar, it can be really interesting working with data because I’ve been with companies that are data hoarders and they don’t realise the cost of keeping all that data and a lot of it is never, ever used. So, how do you prioritise and just understand which important, which data do you want to keep and potentially use later, or in which do you focus on right now? Or are you assigning value to different data types?
Omar Sallam (15:31):
So, this is a question that the answer can be extremely relevant to the scope of type of data or the scope of the products that you are trying to support. As the company grows and as the business cases evolve, the important types of data will start to present themselves and they’re going to present themselves if they’re not really considered from the start, even when things were very simple, they will start to present themselves in the form of problems and issues that are occurring with this specific type of data that are causing challenges and that are causing problems to the end user at some point. Meaning that across your experience, if you started to get really on a big scale, you would be offering inconsistent data, for example, or data that might be contradicting for the users themselves that might start to cause confusion and cause you some problems. And this is the worst that can happen, but that’s also the most obvious way that this specific type of data would start presenting itself as a type of data that needs special consideration in your company or in your product.
The other way around is to be aware of relevant also to your type of product. What is my product and what’s trying to serve? What kind of service or what kind of experience I’m trying to give my customer? And logically, what kind of data would be most important for me? So, for example, in our case in Booking.com, geodata is one of the most important types of data in the essence of the experience that Booking.com is giving. And that’s why it was validated that we need to give some special attention to geodata because we are trying to give more information and high quality of data to the customer to be able to take better decision about planning their trips, for example. And this comes from the, as I told you, the essence of the experience that we’re trying to give.
So, based on your product itself and the business domain that you’re in, you could start focusing on specific types of data relevant to the industry and relevant to the experience. And even if you’re still at an early stage in your company lifecycle, you could start considering not even have a separate product manager for a specific type of data, but at least engineering it and building the architecture around it in a structured way enough that when you scale up in two or three years, it wouldn’t be a total mess and then you would be able to focus on it much easier.
Randy Silver (18:31):
So, that leads on to something really interesting, the distinction between data product versus data as a product. Can you talk a little bit more about that distinction and where we draw that and why it’s important?
Omar Sallam (18:43):
So, this expression is always really confusing because they have different meanings, but as you start talking more and more about data as a product, you get bored of saying the whole thing and it evolves to data product again. So, it’s a little bit confusing, but just think about it as this, a data product is a product, a normal product, front-end experience, whatever, something that you deal with as a user that is powered entirely and mainly by a specific type of data or a specific set of data types. So, a very, very easy example for that is Google Maps. Google Maps is a front-end data product and it’s based on geodata as well, the maps data, the locations, the POIs, the road networks, and everything. And these are managed separately as their own data sets with the teams that are entirely working on that.
And then we have the data product that is the front-end of the user experience of Google Maps. This is a data product based on a dataset that is managed separately. Data as a product is the dataset that is managed separately as I told you. And these teams are working on how to evolve the data, how to manage and serve the data in the best possible way, how to increase the quality of this data, how to build new data structures to serve more complicated use cases, how to build machine learning models that understands the data and generates even higher quality of data and so on. All of this entirely to be able to serve it to some other teams to use it in their front-end experiences like the Google Maps front-end experience. So, this is the difference of the two concepts itself.
Initially when everything was started 10, 15, 20 years ago, you would always have a team that is responsible for the front end and the back end and the data that supports it. And they were responsible for managing everything related to the data itself that I was just listing within the scope of supporting this front-end experience. But as I said, as the scopes get bigger and much more complicated and use cases get much more specific and sophisticated, it presented the need to have a separate consideration and separate scoping for such teams and such products.
Lily Smith (21:19):
You mentioned earlier, Omar, that as the company grows, you might then have that need to hire a specifically data product managers. And before that time, we should just make sure that we’re paying attention to the data that we’re capturing and how we’re capturing it and how we’re using data to ensure that it’s in a good state to be, I guess kind of manipulated and used in the future. How much of a kind of responsibility does that lie with the product managers in the team versus engineering, for example? Would your expectation be that product or product management are kind of taking quite a proactive stance in ensuring that data is managed correctly?
Omar Sallam (22:10):
Yeah, so the responsibility is always shared, but at the very early stages, I would say that most of the responsibility here allows on the technical teams and mainly the people that are responsible for designing the architecture for the data. But that also doesn’t mean that whoever is responsible for product is not involved because on the other side, the product person at an early stage startup, whether even a junior PM, or the COO, or one of the co-founders, they have to be totally aware of the importance of this and establish it as one of the important points that they have to consider and they have to push the technical team always to maintain and to keep. So, putting it as one of the definition of done for each and every task or as the main requirement in designing a new big thing or something like that, it has to be always pushed and considered and highlighted as one of the main priorities when you’re creating a new feature or designing a new product or building something.
But again, the actual design of everything lies on the architect lead or the tech lead that is designing something from scratch. And it would be much easier to design a structure from scratch then to have a big monolith out of control growing for a couple of years and then to fix it. And that’s one of the biggest challenges that everyone is trying to deal with right now, especially for companies that have been scaling for like 15, 20 years and everyone is have this expression trying to get out of the monolith. So, it’s easier now for early-stage startups or even medium size companies that are scale up right now because everyone now realises the complexity of this and to what extent it can get complicated. So, it’s easier to consider it from early starts.
Randy Silver (24:20):
I’ve lost track of the number of companies, larger companies that I’ve been with that have embarked on a data lake, data consolidation exercise things like that to try and pull this all together. I’m curious for when you go into these organisations that are larger that have lots of legacy data that’s in the state that you talked about with some of the challenges. It’s fragmented, it’s got different structures, it’s owned by, it’s very siloed. How do you show value from this kind of approach? What’s an ideal stakeholder look like and how would you work with them?
Omar Sallam (24:55):
So again, this is a very relative question, and actually one of the challenges here is if you’re trying to introduce a new data team, where to place them in your hierarchy and your structure, which department or which business unit they should belong to, how they work with everyone else. Because specifically data teams especially that if they’re being newly introduced, they work really independent from most of the also group or previously defined structures and departments and so on. So, it’s again on the PM and on the leading parties over here to identify who are the stakeholders. And again, communication here, extensive communication here is key because you have to speak to everyone. Think about as a new product manager in a new company, and you have your 30, 60, 90 days plan to onboard yourself within a company. One of the first things that everyone advises you to do is to speak to everyone and try to identify who are your stakeholders, create a stakeholder map, know who’s more important that you who has more key use cases for you and so on.
So again, you are introducing a new concept in your company, so you have to onboard yourself within the structure of your company you have to speak to everyone, literally everyone. Of course, as you understand the type of business and type of industrial company is in, there would be some teams that would be more obvious, that would be more important for you. And so, you target those first and then you start digging around the structure and trying to relate and connect yourself to different teams and speak to them. And after when you do that, you would start step by step proving your worth and proving your impact and establishing that with different teams. And hence you would start creating these impact funnels and grouping the different impact that you’re trying to proxy through the different teams.
So again, it’s an extensive communication process. It doesn’t come easily, it doesn’t come overnight. It’s a long process. But the importance of managing this to avoid having complexities and problems due to data mismanagement is really worth doing this whether you already realise it because the problems started to appear or you’re trying to make account for it. And at an early stage and as management, you are willing to invest time in order to establish this and give it time to sink in and to yield its impact and benefit.
Lily Smith (27:51):
So, what does your roadmap look like? Is it full of outcomes or what do you have on your roadmap? Is it a list of projects for that you are doing to support other teams? Or…
Omar Sallam (28:06):
Yeah, so that’s the tricky part because always on… As a product manager that there’s significant amount of effort and significant amount of the work that you do is project management. But in data products, this volume of project management is a little bit higher because again, you might start a couple of initiatives and a couple of many projects here and there that are aiming to even give you the space to discover what kind of impact that you drive. So, not from day one, you would be a hundred percent clear on what are all the metrics that you are trying to optimise for and the kind of impact that you’re trying to achieve. Because as we establish now, it’s not easy to define this kind of impact and you’re trying to proxy that through other teams and you are working on something that is, it’s extremely technical and extremely infrastructure.
So, you will always find things to work on that are just infrastructure. It’s not easy to drive impact from those. So, it’d be running a lot of various sizes of projects within developing your product itself, but you need to find the value that you are trying to achieve anyways. So, if you still don’t have a direct impact that you are trying to give to an end user, at least the project that you’re working on is going to elevate your value proposition somehow to be able to discover a new area that was untapped for you or undiscoverable for you before and that might unlock some impact for you. And by doing this continuously and building all of these things, connecting them together and building them on top of each other incrementally, you would start realising this state of impact that you are trying to achieve somehow.
So, it’s not always straightforward what kind of metrics? It’s not like managing a product that is customer facing and you have a straightforward dashboards of clicks or conversions or orders or something like that. It’s not straightforward like that. It takes time, it takes effort and investment and actual belief to work on this and reaching it at some point. So, depending on which stage your dashboard, or your map would totally look different depending on which stage you’re at and what are you trying to track or achieve right now. Are you doing more project management or more product management and how everything is tied together to reach some kind of impact at the end.
Lily Smith (31:11):
Omar, thank you so much. We’re out of time that’s flown by so fast, but it’s been really great hearing about your experience as a data product manager. Thank you for joining us.
Omar Sallam (31:23):
Thank you so much for inviting me and having me. I really enjoyed this discussion.
Lily Smith (31:37):
The product experience is the first.
Randy Silver (31:39):
And the best.
Lily Smith (31:41):
Podcast from Mind the Product, our hosts are me, Lily Smith.
Randy Silver (31:46):
And me, Randy Silver.
Lily Smith (31:48):
Louron Pratt is our producer and Luke Smith is our editor.
Randy Silver (31:52):
Our theme music is from Hamburg based band PAU. That’s P-A-U. Thanks to Arne Kittler who curates both ProductTank and MTP engage in Hamburg and who also plays bass in the band for letting us use their music. You can connect with your local product community via ProductTank, regular free meetups in over 200 cities worldwide.
Lily Smith (32:13):
If there’s not one near you, maybe you should think about starting one. To find out more, go to mindtheproduct.com/producttank.