Dynamic pricing is a tool, yes, and while it's an insanely smart tool, it can only work with the information it's given. What you put into a dynamic pricing solution as input matters and makes a huge impact on the price advices it creates. If you have bad competitor data that isn't up-to-date, for example, then the tool will generate equally bad price advices.

[00:00:10] - Grace

Hello and welcome to Price Points, the podcast that examines the changing world of e-commerce one episode at a time. I'm your host Grace Baldwin. And today we're talking about data quality.

So this is a theme we're going to explore this whole month because it's super important. Dynamic pricing is a tool yes, and while it's an insanely smart tool, it can only work with the information that it's given. What you put into a dynamic pricing solution as input matters and makes a huge impact on the price advices it creates. If you have bad competitor data that isn't up to date for example, then the tool will generate equally bad price advices so in the coming weeks, we'll see a lot of content from us about this. But today I sat down with Jasper Wiercx, one of our Solution Consultants here, to talk about data and why it's important. You've already met Jasper in previous episodes and I'll include links to some of his other interviews in the show notes.

In this episode though , you'll learn about what kind of data we need, how the dynamic pricing tool actually collects that data, what a data quality assurance process looks like, why data seems to be so expensive and more. So I'm just going to keep this in short and let's just go ahead and jump right into it. Please enjoy this interview with Jasper Wiercx.

Well, so thank you again for joining me. I wanted to start a little bit talking just about data quality in general. And so, data has become more important in every industry. And so why is data important in pricing and what kind of data is important?

[00:01:41] - Jasper

Yeah. So I think in general like you said as a company in general or an organization in general, I think you can definitely see a trend into the digital adoption.

And if you want to have some way to perform in [00:02:00] a digital space, you need to have data available to position yourself, but also to make certain it's to gain insights out of the data and become more of an insights-driven organization. So as a retailer or as a brand moving into Dynamic pricing, of course, you'll be very dependent on the data that you use to base your decisions on. In that regard the quality of the data is very important, of course because you make automated decisions and at least I would be only would be confident about those decisions if I'm a hundred percent confident that the data on the line those decisions is actually accurate

[00:02:48] - Grace

The data that we're talking about is pricing data from both competitors and then comparison shopping engines, correct.

[00:02:55] - Jasper

In terms of external data. Yes. So about the comparative data or retailer data in that respect is primarily pricing data, but you can also argue that your internal data should be accurate so. Do you want to have includes more advanced Dynamic pricing strategies like stock-based you need to be sure that those data elements as your stock units sold for instance are also accurate because otherwise you will still make wrong decisions based on flawed internal data.

But what if you sell your own products or have products that can't be matched, can't be compared in a one to one comparison. Then what can you do? You need different data points to steer on, to base your price on, and one of those is reference pricing. So what you do is basically for a product that's hard to compare, you make it comparable. You refer the price of one product to the price of another product that's not directly comparable.

In short that's reference pricing it's used a lot in private label products for instance but can also be used in matchable products or unmatched products, but for private label it's it's a very good one because it also places you in terms of value in a relative distance towards known products. Let's say I'm making my own TV. I know, yeah quality is about 20 percent difference or 20 percent less then I'd like my price also to be 20 percent less than a famous Samsung TV for instance. So I'd refer both in quality and in price towards other other products and thereby making it comparable.

[00:02:48] - Grace

The data that we're talking about is pricing data from both competitors and then comparison shopping engines, correct

[00:02:55] - Jasper

In terms of external data. Yes. So about the comparative data or retailer data in that respect is primarily pricing data, but you can also argue that your internal data should be accurate so. Do you want to have includes more advanced Dynamic pricing strategies like stock-based you need to be sure that those data elements as your stock units sold for instance are also accurate because otherwise you will still make wrong decisions based on flawed internal data.

[00:03:30] - Grace

Yeah, and as they say garbage in garbage out, correct.

[00:03:32] - Jasper

Yeah. No exactly. Yes. That's the typical way to put it. Yeah.

[00:03:37] - Grace

So the crawler is separate from the dynamic pricing tool the crawler collects the data. Yeah, then whatever Dynamic pricing company you work with will check the data and then feed that data into the dynamic pricing tool, correct? That's the flow of logic.

[00:03:50] - Jasper

Yeah. So, yeah, actually you should see the exactly as separate. It's the fact that you have a dynamic pricing so far which automates certain decision-making on data, basically. However, the data needs to come from somewhere internal data is of course obvious. You will provide it to the software. However, the external data or a competitor and pricing data that something at the crawler looks for and makes available to that software to make decisions on. Scraping information online is just a very popular thing to have but that also makes it very difficult

[00:04:27] - Grace

Popular in what sense so popular for other dynamic.

[00:04:30] - Jasper

In the high in demand, I have to say.

And it makes sense because it gives you a clear insight in how a competitor or a comparison shopping site in that case is actually offering the same type of products as you are offering. So if you can get insights in though that information so how does a product get positioned online but also at what price does it get offered that will give you Insight in how a competitor is actually operating the also means that a lot of competitors they take action to make it more difficult to scrape that the information which makes scraping and crawling websites difficult thing to do.

[00:05:15] - Grace

So what sort of actions do they take to prevent you from gathering that data?

[00:05:19] - Jasper

Yeah. So that's a good question. If if you are a crawler you often use an IP address to log into a website and if that website detects you entering that's websites too frequently from a single IP address, it will notice that and it will block that IP address. So as a crawler, you probably have multiple IP addresses or multiple servers with that via which routes your crawler can actually go to the website in order to get that information.

And those got this is an example of some kind of constraints or some kind of [00:06:00] barrier. That's the this gets competitive retailers take in order to limit access to the information. So as I know is for instance Amazon as a good example is one of the most difficult players to actually scrape because yeah, they don't want to get that information widely available for everyone.

So if for any reason a competitor or for instance Google Shopping as we've seen a few weeks ago actually changes the layout of its web page, that will also impact your crawler and that's what of course impact the data quality for scraping in general. You're always dependent on what is displayed on the on the website.

Yeah, and as you also mentioned, we are we either of comparison shopping sites or we directly straight from the website. Google Shopping, for instance, does not display the stock levels from the retailer, so we cannot all we cannot give you information on the stock level if you have Google Shopping as a source. If you directly scrape, we can include of course the stock levels as we will scrape that.

[00:07:41] - Grace

So that's a good transition that into talking about dependencies.

So. Why is it important than to have data from multiple places for that is that's a good example, I guess, right.

[00:07:52] - Jasper

So this first of all it's an indeed about what type of information is readily available on the source like we discussed with stock levels. Secondly, it's also as we just also briefly talked about is the fact that the web page is websites are constantly changing.

So there by if you only have one data source, you will always be at risk of changes in the data source, if you combine multiple data sources you mitigate your exposure to the risk of someone else changing their their their layouts for instance so

[00:08:34] - Grace

So is then one of those examples so we're talking talking about GTIN versus URL.

So we use both of them and that's a way to reduce that dependency on a URL breaking.

[00:08:46] - Jasper

Yeah so, not necessarily URL breaking. So we always look for the ID so the GTIN for instance or the EAN codes on a URL, okay. If the  URL changes, then we of course need to inform the crawler or update the crawler on the fact that that URL actually changes.

So as an example, let's take let's take Amazon and Amazon for whatever reason changes its URL our, crawler cannot find that URL any longer. So therefore it cannot find that set of GTINs to be scraped from Amazon. If you know also have a Google Shopping as a data source where Amazon actively promotes or sells products, you will still find that set of cheating products on Google shopping. Assuming that that's that's Amazon selling that via Google Shopping.

[00:09:46] - Grace

And so will that price on Google Shopping always be the same price that's on the website then.

[00:09:51] - Jasper

Yes. Yeah, yes it because school shopping requires effect that your the price that they offered a product for via Google Shopping is accurate

[00:10:01] - Grace

Okay. Yeah, and so if for example Amazon isn't selling on Google Shopping or isn't promoting on Google Shopping, is there a way for us to detect whether the website has changed or URL is not working?Are there any categories that this is especially useful for? And are there categories where it's not useful?

[00:10:13] - Jasper

Yeah. So so of course, we have a large client base all tapping into different data sources.

So if anything happens with one of those data sources, we know it quite fast. And we inform our other clients who were not aware of the change and how it might impact our data quality and data accuracy. Also, we have other mitigations in place where we actually store accurate information for extended periods in order to mitigate these type of risks because we are also dependent on that. But thankfully we're working with a few very professional partners that notify us more in advance that we are very happy with in terms of data quality and we are completely [00:11:00] confident that they will do their utmost best to change their crawler as fast as possible and again update that pricing quality.

We make the selection quite carefully before which partners we are working with both in terms of their professional. But also in terms if they if the data is actually accurate and it's a good that it's a good quality of data because there's a lot of scrapers out there offering rubbish data.

It makes sense isn't it data is important for organizations so people are trying to monetize on that and then trying to sell the data. In that regard. you see the value of data actually going up and therefore all often also the price and those enterpreneurs or trying to take of course advantage of that by selling the same data or perceived same data the lower price. But there's also reasons why of course a price of data is at a certain level [00:12:00] is because the quality needs to be the best and therefore we decide to work only with selected select the scrapers of course and more importantly is that we as only are we offer a One-Stop shop of different data sources of those premium scrapers.

If you for instance develop your own dynamic pricing solution, you would actually have to procure the data from those individual scrapers and then again combine it in your own system. And we can actually easily tell you which data source is actually the best for your specific situation.

So that means determining which data source has the best match rates on your competitors in a certain market, but also to only select certain [00:13:00] top seller products of your own. And directly scrape different sources from your from the your most important retailers to ensure that you have the best coverage of the data

Yeah.

[00:13:14] - Grace

So is there like sort of a rigorous test that that a partner goes through before they become one of our partners?

[00:13:20] - Jasper

Yeah. No, absolutely. So both in terms of the process both in terms of customer service, of course, and also in terms of just the quality itself to we personally check very very well before actually selling it or offering to our to our own clients

[00:13:37] - Grace

Okay. Yeah, and so if someone you know is dead set on building their own, are there any quality checks or what would advice would you give to someone who wants to actually procure their own data and work with data sources on their own? How can you avoid the ones that are selling you snake oil? Or is it really hard to tell?

[00:13:59] - Jasper

Yeah, I would talk to a lot of different layers, of course and just do a very solid data check on it before you make any decisions. I've seen from multiple clients that they already had a certain certain provider in place, but that the data that we provide was actually more accurate. And for some reason all the active users is already knew that the data was pouring some sense and it also it also makes it difficult to completely trust the system itself, but how to check how to do that. Yeah, there are multiple ways to do it. And if you want more information, I think you just we can do a check for you if you are curious and see if it

[00:14:47] - Grace

Yeah. Yeah get in touch

[00:14:48] - Jasper

Exactly

[00:14:49] - Grace

Actually talk to somebody who knows the different data sources available

[00:14:54] - Jasper

Nowadays, I think as a retailer or as a brand you want to sell your products, [00:15:00] that's what you're good at. Do you want to position yourself correctly in marketing? There's no reason to spend a lot of time in these type of nitty-gritty details, but very important details, without really knowing what to what to look for. Yeah. Yeah, exactly for all the. Let's say just a few dollars in the penny in on the monthly monthly cost.

If you would compare that to the actual time effort you would need to spend in it. You could very much argue that your time is way worth to spend on something else.

[00:15:35] - Grace

Yeah, because I mean we have a full team here doing quality checks and everything. Yeah, exactly. So an organization or a company that wants to really do this themselves would need to hire a ton of people to do this yeah, and then take a lot way a lot of resources from their Core Business correctly.

[00:15:50] - Jasper

And value based pricing starts off then with mapping out your assortment based on some characteristics.

[00:15:52] - Grace

cool. All right.

[00:15:53] - Jasper

Just one more thing. I think it's good to mention. So. [00:16:00] We offer the data ourselves because we feel like it's important to offer that as this the only way that you can use our system or dynamic pricing system in the case, but we also offer the solution that you can give us your data and we plug it into our system. If you really want to.

[00:16:19] - Grace

So why would someone want to do that? Oh, so you mean so someone like a company can come in and say hey we are we're using this data source. We'd like to install it into a dynamic pricing solution. Can we do that?

[00:16:31] - Jasper

Yeah. So let's say if you already have a data provider and you might be dependent on a certain contract term that you have when you don't want to double up on cost and I can completely understand that.

And then you can we can use that data plug it into our system and then you can still use dynamic pricing. Simply for the reason that we feel like if you really want to we want to give you that flexibility and if you really think that's going to make your business successful we're willing to support that.

Of course, we are going to challenge you on the data accuracy simply for the reason that you will not experience a good dynamic pricing solution or our tool be very much dependent on the data that you give up to us.

[00:17:18] - Grace

If a company is coming in and they already have a data provider can they use that and then also enrich it with data that we have

[00:17:24] - Jasper

Absolutely.

[00:17:26] - Grace

You can get like a quadruple whammy there? Yeah, actually from a ton of different data providers. Yeah, which sounds like the kind of the best of both worlds if you're in a contractual a contractual situation.

[00:17:37] - Jasper

For instance or simply because you just want to look into whether how dynamic pricing would look but to be to be fair, we've also seen in the past that the actual results of dynamic pricing was quite was quite bad because the data underlying was just bad.

[00:17:57] - Grace

Yeah. Yeah makes sense Perfect. All right, well. Perfect. Alright. Well, thank you so much. This has been really helpful. It's definitely upended a couple of notes I had but for that's for the better. If people want to get in touch with you how just by email and then also LinkedIn is okay.

[00:18:15] - Jasper

Yeah, so LinkedIn is okay and my email is Jasper at Omnia Retail.com or reach out to via the website.

[00:18:22] - Grace

All right, perfect, and I will include all of that in the show notes. Yeah, thank you.

[00:18:26] - Jasper

Thank you so much

[00:18:31] - Grace

Thanks for listening to price points. I hope you found this episode insightful and learned more about data than you expected. If you'd like to chat further about data. You can reach out directly to Jasper at Jasper at Omnia retail.com. That's J-A-S-P-E-R at Omnia retail.com or via LinkedIn. If you'd like to chat with me, you can also catch me via the same channels.

In the meantime, though I hope you have an awesome rest of your day.

 

 

 

SHOW NOTES:

Omnia was founded in 2015 with one goal in mind: to help retailers take care of their assortments and grow profitably with technology. Today, our full suite of automation tools help retailers save time on tedious work, take control of retail their assortment, and build more profitable pricing and marketing strategies. Omnia serves more than 100 leading retailers, including Decathlon, Tennis Point, Bol.com, Wehkamp, de Bijenkorf, and Feelunique. For her clients, Omnia scans and analyzes more than 500 million price points and makes more than 7 million price adjustments daily.

Website  LinkedIn

Music: "Little Wolf" courtesy of Wistia

TO CONTACT HIDDE ROELOFFS VALK:

Email: hidde@omniaretail.com
LinkedIn: Visit here


TO CONTACT GRACE BALDWIN:

Email: grace@omniaretail.com
LinkedIn: Visit here