Price Points Podcast EP 5: Building vs. Buying a Dynamic Pricing Solution

[00:00:10.180] - Grace

Hello and welcome to price point the podcast presented by Omnia Retail. I'm your host Grace Baldwin. And today we're talking about what it takes to build your own dynamic pricing solution. Something we hear a lot especially from larger companies is that they want to build their own dynamic pricing solution in-house. And the reasons why are understandable. You want to keep your data as centralized as possible. And if you want to build your own solution that's totally fine. But you should prepare yourself for success what goes into dynamic pricing solution from a technical perspective and what sort of manpower do you need.

I ask these questions and many more to Berend van Niekerk our product manager here. Berend is hard to describe. He's quite a character. He's absolutely brilliant. But his mind moves faster than his mouth sometimes. And what's funny is that he's the first to admit this. He has a master's degree in industrial engineering and management science from and Dutch speakers you'll have to excuse me here the University of Groningen. And he has been with Omnia from the start. Today as Product Manager he is responsible for the algorithms and the functional development of the software gathering all the e-commerce and software knowledge within the company, combining this with the wishes of our leading e-commerce customers, and translating this into a tool that helps customers automate their marketing and pricing strategies.

Berend took the time to explain to me what actually goes into making a dynamic pricing solution powerful and gives you some practical tips on building a solution that works for your company. So if you're considering building a solution for yourself you're going to enjoy this one. So please help me welcome Berend van Niekerk.

So let's get started. Thank you for sitting down with me. Berend can you tell me a little bit more about yourself and what you do here at Omnia. How long have you been here.

[00:01:59.650] - Berend

Sure. Yeah my name is Berend van Niekerk. I started only almost six years ago so almost from the start. I've been involved in a whole growth path. We we went through the beginning we were just a small company providing data insights in pricing a bit of pricing automation and evolved over times into the full company where we work for bigger enterprises now actually serving a full suite of both pricing and marketing has been very excited.

At the beginning I was involved in everything from customer service to consulting to sales and a bit of further management of course. But nowadays I am responsible for these functional developments at Omnia, so meaning what does the system do in actually translating our own customer needs to a roadmap and do that together with our Chief Architect as I'm responsible for what it can do and he is responsible and how it's actually done. It's very exciting and a lot of developments going on an exciting future.

[00:03:00.560] - Grace

So your actually wrote Product Manager is your actual title.

And so you're just really responsible for setting the roadmap figuring out where we're gonna go and then Nathan who's the Chief Architect is more responsible for how we're actually going to do that right?

[00:03:16.540] - Berend

Yeah yeah basically he's responsible for the execution so that the code around it the database is not responsible what the system does. So do you need to calculate a sort of thing. Do I need to have an interface around it. How does it need to look. What does it need to do. What kind of business rules are in the system and how can we improve that our customers can do to work easier and make their everyday lives that company happier.

[00:03:40.720] - Grace

Yeah makes sense. And so today we want to talk about building versus buying because that's something that we see a lot in the in the market is that some people say you know we're working on building our own solution we don't really necessarily want to buy dynamic pricing software when we can do it in-house. And so I wanted to explore that a little bit with you and talk about what are the pros and cons of both building a solution but also buying one. So maybe it's a good place to start to talk about what actually goes into a dynamic pricing solution. And like how that how you can materialize that by yourself or with a company.

[00:04:19.510] - Berend

Yeah I think it's a it's a very important part of the onboarding and this is of course not only the alignment between the business and I.T. but also within the business there is various departments that need to be aligned on what you're trying to achieve with your dynamic pricing.

So I think at the beginning customers think that's OK I need to have a data set so I need to know what the competitive pricing and market are. Then I need some basic pricing rules which is built by a BI guy or by a guy from IT and I don't need to upload it to my system. And then I got a full dynamic pricing system. Um in reality and those three parts so the data, the pricing logic, and actually the automation around it to make that a success, there's way more around it. And to make it fully adopted by a company and I think that tradeoff is often not made at the beginning but companies start to realize that once they go into adoption they think OK to make this a success, it's not as simple. So I need to build certain systems around it. I need to have the knowledge in-house. I need to further develop it. I need to maintain it. And what you often see happening is once they try it out and they come back to us again because we've got a full suite solution.

[00:05:35.650] - Grace

Can we talk a little bit so like what sort of systems do you actually need.

[00:05:38.740] - Berend

It all starts with the data actually. So you need to have a full overview of the markets and up to date overview to make great decisions.

That sounds easier than it is. So what you need to do you need to have connections with multiple data suppliers to provide you the full overview of who is selling what against what price with what delivery times and means you get it at continuously and actually update that multiple times in day. So that's more to market side. You need to understand what is happening who's selling for what. And the other side of the data you need to update is your own data. You need to know OK what are my exact marginal costs.

So it was my purchase price. What are my delivery costs. The logistical costs what are my marketing costs to sell a product and you need to have all the product information around it. So what are certain categories. What is my lifecycle over product all that information need to get out. They need to store somewhere and update continuously

[00:06:37.540] - Grace

Does that information change a lot? That internal information?

[00:06:39.580] - Berend

Yeah both so internal information categories for example of a product don't change as much but the purchase price costs around it, they are often renegotiated and we need to feed it towards us and also to the stock levels for example and stock levels not just how much you are gonna stock but what you often see at our customers is that they got different warehouses and different warehouses with different stock levels and those different warehouses mean that their delivery time is different and based upon that. It means also that you get longer delivery times you're less competitive versus your competitors so you probably need to make another tradeoff.

So actually that continues. That always updates you need to actually update at least once a day but preferably multiple times a day. You need to get to fresh information to make the right choices at that time.

[00:07:30.230] - Grace

And is that pretty. That sounds pretty labor intensive to make those updates right.

[00:07:34.790] - Berend

Yeah it is a process that you need to automate. I've seen it happen sometimes that it uploaded had somewhere manually but that's not enough to automate a poem.

So what you want is a full live connection that is set up once and easy easy to change also. So if you need to have additional fields from a from your own database it should be able to add it to the system. Sure you can make the right choices upon it. So example can be a customer of ours. They feed their life cycles and they got a different rule in that. So you want to add a field to the feed easily connected to our system and then make sure that you can use it later.

So for example if it's in lifecycle one which is a new product you might want to be in a bit more aggressive to to let the market know you're out there. But if it ends and you don't have enough stock anymore more to lost products you probably want to have a different pricing route. So all that customer specific dynamics you want to capture within the system to allow to put rules on that and to automate eventually on it.

[00:08:37.390] - Grace

And so what about data validation. How important is that?

[00:08:41.810] - Berend

Very important especially in automation. Data validation is if you if you got to manual step in between somewhere.

So if you got a system that you apply some data you do some check you do an analysis and then you get some prize advices and you manually check them and upload it to the system. You actually got manual steps going in between which you can verify does this system do the right thing and then you can upload it to your own system. Unfortunately nowadays dynamics they go way faster so meaning that you need to change your price multiple times a day over your complete assortment to be able to do that you don't go into manual checks in between.

So you want to fully automate it and you want to trust the system on it. So you want to actually data all the data that comes in is a certain data validation. So we make the right choice every time and you can then automatically change your price to it without doing any manual checks in between. That means you need to build a whole system around it that supports that the dusted checks and that has the right feel safe in place and eventually that you feel comfortable automating the practice of your full assortment within maybe a couple of boundaries but you want those boundaries to be as wide as possible.

So you actually automate the biggest part of your assortment continuously which is a quite a challenge.

[00:10:05.210] - Grace

Yeah so I was going to ask. So that sounds that sounds like it's I mean how much manual labor do you think goes into just building that kind of a system.

[00:10:13.550] - Berend

There is it's all often I think underestimate it.

Um building a complete system with the right failsafes in place takes probably a year's worth or two to do completely. There's not only data validation but is also the validation in later steps in the system so you get the validation of the data which is the input then you've got certain rules you apply and you want to change those rules continuously. So any change you do in those rules. Every little change you want to allow the end user to do that immediately but that means you don't have much time to check all validations upon it.

So the system itself should due to changes within the developed nations within the system so you can change the setting. It's validated immediately that it's correct and does the right thing and then you can automate upon it. So it's a it's even more than just data validation you also need to validate all the settings around it and all the changes you do.

[00:11:15.920] - Grace

Yeah that sounds like an undertaking.

[00:11:19.260] - Berend

Yeah.

[00:11:20.010] - Grace

Um so moving into the the pricing logic. So can you explain that a little bit in the pricing rules and what. So if someone wants to build their own system, what do they need to know about pricing rules and pricing logic

[00:11:32.630] - Berend

Yeah. So what you often see is you've got a dataset of all your competitors so you know who your competitors are selling the same products. You can start making choices upon that. There are different level of choices. So what you can do you can follow a certain competitor and saying OK I want to follow competitor A and change my price which is which is quite simple. You can also follow a group of competitors at the cheapest price of a certain competitor. You can also grow a bit more advanced and say given the past performance if I follow certain competitor then I see this happening.

It means I need to have a certain distance of that competitor to have an optimal price. So that's more towards intelligence. You can build that.

But what it means you need to think about the whole strategy upfront and then you need to ask your I.T. department or your BI team to build all those rules and what you don't know most of the time in advance is how will certain things react in the market. How will your competitors respond to your changes? So what we often see happening if if our customers build their own system is that they want to think about everything that's going to happen in advance and they want to implement it.

And that means that it's a very big change you want to do at once which is scary. You want to validate that and it's difficult to implement what you actually want to allow to use it to do in those rules is like also I think Travis mentioned in his last podcast is you want to use your to take baby steps and iterate all the time.

[00:13:12.630] - Grace

Side note: Berend is referencing a recent episode I did with Travis Rice one of our customer success managers. I'll link to that posting that description below. If you're curious it has price points Episode 3

[00:13:25.290] - Berend

So do a small change first maybe take a part of the assortment. Do a simple strategy change implement let's see what the results are. Learn from it and change it again. If you want to allow users to do that you need to have a system not only pricing rules but actually a full interface around it which provides users the insights in the market so understanding which other competitors that are selling my products. How do you response to changes. I do. What are my margins I make on each price level.

So if I follow a competitor A Do I still have enough margin left to make a good profit. All those inside you need to capture within the system. So you need to have a portal around it to provide your insights from those insights. You can set new rules and new rules. Indeed you can put in your own I.T. system what you want you actually want to allow is the end user to each time he he thinks of opportunity big or small you want him to be able to execute that immediately. There should be no barrier to changing that rule because if you lowered a barrier it means they can experience more they can build out their systems step by step and eventually you will have a way better pricing system.

[00:14:42.310] - Grace

Sorry to interrupt but so that you also want to be able to do this on a product level. Right?

[00:14:46.810] - Berend

Yeah. Yeah. So you want actually to do it on every cross-section of your assortment which is important to you. Sure it can be simple things like categories or brands which are fixed subsets of your assortment but it can also be on product level if you really want to do it. We most of the time digital files to go to product level changes because that goes against automation. So what you actually want to do you want to think about okay Why do I want to change this specific products to a certain price.

Often there's a logic behind it. For example the product is in a certain promotion so I need to fix that price for a certain period.

What you what you can do is get all the product ideas and set a fixed price to it about what you actually want the system to do is recognize that it's a promotion for all the product that are in promotion just used to promotion price. That means each time a party goes in promotion automatically the system changes to that price. That means it saves you tons of labor and you can build your pricing strategy piece by piece and eventually focus on the strategy instead of the execution which is which is key and you want to have a system that allows you to do that.

[00:16:01.830] - Grace

And does Omnia and let you do that?

[00:16:03.160] - Berend

Yeah. Yeah. So the promotions is an example. The lifecycle for products I gave earlier is another example.

But also what you can do you can import any viable that you provide into a feed or into a connection to own the app. So there are a few fixed variables like to get original brands more dynamic for rivals like a promotion.

We're still quite static but you can also go way more advanced and thinking okay what is my stock level that continuously changes and what are my skills over the past weeks. And if I got too many weeks of sales left I might want to go more aggressive so then it becomes more dynamic and within Omnia you can connect each field quite easily so once you got to connect set up you get a list of fields and you say Okay I want to use that we've been on here you can alter it a bit if you want to should do a calculation opponents and then we import it multiple times a day with the latest values and then based upon all the changes that are happening during a day we can set automated rules so that your strategy is being executed continuously to the to the product level.

[00:17:15.160] - Grace

So then once you have the price advices you will want to be able to put them in a store right.

[00:17:20.620] - Berend

Yeah you want to.

Once you have calculated the new prices you want to upload them to your e-commerce platform where you display them on your website or even third area you want to also change them within the stores itself. So what you see happening more and more is that retailers have electronic shelf labels to be able to cope with the dynamics of the online market of the online competitors I must say.

So you need to allow the system to do it continuously and you once to not have any manual steps in between. So you want to have the system calculate everything from the data to the logic and then providing prices back as the market changes continuously during a day you need to change prices at least once or multiple times a day. That means you can't have any manual steps in between.

So that means you need to automate fully. Being able to automate fully, you need to validate all the steps in between you need to know you're safe. So what you want to do is you want to make sure that within certain boundaries you can have loads automatically and fully and can trust the system to do so. That is a big challenge. So that's the whole chain from the data input you getting to where you export you need to have four steps in between validating data validating calculations should do. Validating that export is right versus what you did before and then you say OK it's within a certain boundary within a certain safety I will love that change to go on that for in-store infrastructure that takes a lot of time to build.

So you need to invest that time. Otherwise you can't automate and if you can't automate you can't keep up with the markets and you won't have the benefits from a dynamic pricing system. So you need to do that for your full assortment continuously. So it's the infrastructure around it which is important. And on top you need to have the raw calculation power because well to update everything to get all the newest competitor data prices you calculate that to the newest price advices and to export at a low of processing power going on so you also need to invest in the infrastructure around it to support that. And eventually if you have those those both you can upload it continuously to your system and make sure that once something changes.

So either your own internal purchase price changes. That means there's a new optimal price or if a competitor changes the price you want to capture those signals and then upload immediately back to your system. The newest optimal price so you can keep up with the dynamics of the market and get the most out of it. The dynamic pricing system.

[00:20:08.170] - Grace

Do you think it's worth it for enterprise companies to try and build their own dynamic pricing system?

[00:20:14.740] - Berend

It's always a trade off you need to make. I would say 9 out of 10 times it's the work that goes into is underestimated.

We got 7 years experience both in building a system but also in the knowledge around it was needed to execute a right pricing strategy. That's a lot of head start we have to for our own companies to build something like that is on the same level we do.

That means you need to invest a lot of resources in it both in the I.T. side as well as on the more the business side so what we need to do what's the business logic. How are the markets reacting. Well the I.T.. It's building the first set up but also maintaining it and doing all the iterations upon it and building the system around it, so a portal and a few full execution.

That's a lot of work. And while every company has is the I.T. resources are really scarce. So then it's a question do you want to invest for five developers full time building dynamic pricing system or do you want to invest those guys into building your e-commerce platform or well anything that's important for you or for your everyday sales. Often that's right of goes to building your own platform and if you use a third party too it means you get more developers left on building the things that are important for everyday sales and leaving the bring pricing to a third party tool.

And on top a tool like ours allows end users like the category managers or the marketeers to use the system themself without any I.T. interference. And so that means you you can be more powerful so every idea you have you can execute immediately and having that I.T. boundary there it means that you want iterate as fast and you want won't build the full strategy you want it to to be to make it a success.

[00:22:20.530] - Grace

So for the final question is there, if people are really convinced that they want to build their own system do you have any final advice for them?

[00:22:32.230] - Berend

Yeah I would say be sure that you invest as much time in the portal that supports it as you do in the logic and the data itself.

Without having the right portal an interface around it it will not be adopted by the users within the company and the change management and adoption that is needed within a company.

You won't succeed if you don't allow people to get along on that journey. To understand why you're making certain changes and to encourage them to think about destroying itself and to enable them to be able to change it without any interference of I.T. spend time on that because that's as important as having rights I.T. logic underneath it's allowing the end user to iterate continuously to have the insights it will enable them to do a better job than if they are blocked by any possibilities. So I think that's the key thing we learned over the past years.

You need to take everybody along so it doesn't. So it is not the black box. You have the right insights and you allow the flexibility within the tool so people can actually use it in their everyday lives and improve continuously.

[00:24:00.570] - Grace

All right. Well thank you Berend. This has been super enlightening. If people want to get in touch with you, how can they reach out to you?

Yeah you can always be in touch if you want it know by email. So that's Berend at omnia retail dot com. Or call the office here. I'm happy to um to talk to you and you can always reach out on LinkedIn if you want to.

[00:24:21.850] - Grace

Cool. Perfect thank you Berend.

[00:24:23.840] - Berend

No worries.

[00:24:32.460] - Grace

Thanks for listening to price points and I hope this episode was enlightening. If you'd like to talk with Berend you can reach him via email or via LinkedIn and you can find both of those in the show notes. As always if you'd like to talk to me, you can also see my contact details in the show notes as well or contact us via our web site. But for now have a great rest of your day.