On-Demand Webinar: Workflows in Product Information Management (PIM)

Julia Miksch: Hello and welcome again to our webinar! Today's topic is "Workflows in Product Information Management." How to sustainably optimize workflows and approval processes in the product data processing. Today on the webinar, Oliver Kleinjans, a Business Consultant at akeneo. Hello Oliver! Great to have you with us! And Markus Kettler, Senior Consultant at diva-e. And I'm Julia Miksch. I'll be moderating the webinar today, guiding you through it, and asking the questions afterwards. So, we're ready to get started. Markus, I will turn it over to you and the participants. Have fun and have an informative webinar!

Markus Kettler: So, welcome again from me. Julia, again, quick confirmation that my screen is coming in? We're watching the start slide right now. As I said, it's about the topic of workflows in Product Information Management. And we're looking at so that workflows and approval processes can be optimized. Simply costs can be optimized, and, if necessary, better results can be achieved. In the following, we have briefly set up the agenda for today. First of all, a short round of introductions. We tell two sentences to us, which hold here the lecture and then also gladly later the discussion. Then we'll go into the topic of PIM, in general, to pick everyone up a bit and make sure that we're all on the same basic level. Then it's, in theory, about the topic of workflows, opportunities, and challenges, what to look out for, where there might be stumbling blocks. And then, with Oliver, our practical partner from akeneo, we get valuable insight. And last but not least, there will be a Q&A session, where we will be happy to take up the questions that have been asked in the meantime. So. Let's start with the introductions. I'll take the liberty of starting with myself. Well, I'm a senior consultant at diva-e, with a particular focus on PIM. I've been working in e-commerce for ten years and working in PIM for several years. At diva-e, I support our various customers in designing and implementing optimized PIM processes. And at the same time, I think it's important not to see PIM as an end in itself, but rather that this structuring of data flows serves a larger goal, namely the optimization of real business success. And you always have to keep that in mind when you're working on PIM. So. I would then like to hand it over to Olli for an introduction.

Oliver Kleinjans: Yes, thank you, Markus! My name is Oliver Kleinjans. I'm a Business Consultant at akeneo, and as a consultant, I support our customers and our partners in the implementation of their akeneo projects. Thanks again for the invitation. I am happy to show you an exemplary workflow within our PIM system. We will enrich product data to a whole level, including a validation step, to finally publish the products.

Markus Kettler: Thank you! Then we'll jump right into the topic and, as I said, first take a look at the case of PIM itself. So Product Information Management, what that means. So earlier, if you look at a customer lifecycle like this, you see here on the left side the traditional cycle that the customer goes through like this when they go through a buying process. That is, you either discover a product by chance or search specifically for a product. At one point, you find a specific retailer who also offers the product satisfactorily. So, you agree on a price. Then the purchase takes place. And if this whole process runs reasonably smoothly, if one was satisfied, then one remembers this process. And then you come back to it without any new stimuli coming from the outside to disturb this learned process. Over the last several years, with the advent of digitization, which has become increasingly advanced, various new channels and touchpoints have been added. That customers can be provided with information from everywhere, are inundated, so to speak, but also have many more opportunities on their own to get information. So when you're looking for a product, you instinctively start an Internet search first, maybe find an article. And then you go back and read different product reviews, testimonials, and maybe watch one or two YouTube tutorials or videos about how the product performs in real life. And then, at some point, you reach a point where you make the purchase. And at this point, consolidated and well-structured product data plays an essential role in really providing customers with the information they need at these touchpoints, whether they are mobile or on a desktop computer. And that then also goes beyond the Internet. Point of sale, directly in the store. That we always have consistent data and at the same time not only good product information that is correct, but also to be able to answer the customer's questions about the product directly. This brings up the keyword of service. We can provide accompanying information now, that the customer can quickly find operating instructions, explanatory videos, news, the complete information package. And when the customer feels well looked after before the purchase, during the purchase. If the investment runs smoothly, if the purchase process is well thought out, and if the customer continues to find the correct information after the purchase and feels well looked after, this customer will become a loyal buyer and come back again and again. The potential use of PIM is illustrated here.

What does that in itself make this PIM system? That is illustrated quite well here. We have, right away, when you talk about PIM, about Product Information Management. At this point, we are already just on the subject of workflows, actually also an extensive data workflow, as you can see here. Or a data flow, much more. We have the area on the left side where the data is collected from EAP systems of suppliers, who provide the data. Different media, different departments. The marketing department includes marketing texts in various formats, i.e. Excel and Word. Often, data is provided in PDF format, which may then have to be adapted. This is then collected in a central location, brought together via mapping and into a uniform structure. So that here, in this blue circle, we do have all the information in one place. The content is always up to date, and all communication is always available. Of course, this includes text information and, if necessary, attributes, characteristics, and everything needed. That media are also linked or stored here. So images, videos, PDF, text files. In other words, everything that plays a role in a product. And that we then also have this prepared data available centrally for translation. The appropriate classification for different channels can be made, and then just on the right side can be distributed in the other media. It goes via its website, i.e., the homepage, where the product portfolio is presented well. All information is easy to find in e-commerce, online stores, and different sales platforms, even on mobile. You can also see that up here. We also have the point of vocal assistant or voice search, which is becoming more and more popular. People no longer search as users by entering keystrokes but search for the relevant information on their cell phones using their voice and expect an answer directly. Exactly. But also beyond that, as I said, the point of sales, social media, social shopping, Google Shopping, all of that is supplied with the central information. And that is what the PIM does. And thus actually offers a single point of proof that keeps all data always available in the right quality. Here is a summary of the advantages of PIM on one slide.

I want to emphasize the blue and the red point here. So above all, simply the increase in success, yes, through the better information situation, the customer receives information he needs or is looking for, but also the product, which then belongs to the correct information. Or rather, the right information about the product that is also ordered. On the one hand, this increases conversion rates. On the other hand, the bounce rates drop when the right product is delivered, which I expect. On the cost side, the return rates drop, recourse claims go down, and we optimize communication with the customer. And at the same time, because we can use the PIM to drive forward all this data preparation and optimization and bring it into a structure, we also have much leaner communication when working with translation agencies. We have no effort here. Everything can run more smoothly. If necessary, we have fewer correction runs and fewer back and forth if everything is well structured. This means that the topic of "workflows" is also directly involved here. Workflows are simply a central component for optimizing product data flows so that you know who has to do what and when. So much for the theory. So much for the approach on the subject of PIM.

Now we come to the theory focusing on the topic "workflows". First, of course, the opportunities arise when you have well-structured workflows. But also the challenges that have to be mastered. What you might have to watch out for. Let's take a look at that now. What are workflows in the PIM area? How can you imagine them? I have brought along a relatively structured workflow. This is an excellent example. This means that we do have a transparent process instead of running in an uncoordinated way. That is, we always have the product created at the beginning. So, the product is possibly born in the EAP system, receives an article number, is then imported into the PIM, and is enriched with data and created and refined. Then, the second step goes toward data verification, where the next person possibly looks: Is all the information correct? Is something still missing? If necessary, this person then suggests changes or requests changes. And if everything is in order, the product can be released. We have already completed the initial work for one or more products. From there, it could then go to product management. That is, to the product house management and then, if necessary, to external use. For display on the website, output in stores, print media, customer catalogues. Anywhere where customers can see and use the data. At the same time, when the product has reached this status, another workflow can be triggered in parallel—simply following this arrow, where we then end up in a translation. The translation agency or the internal translation department employees who translate the texts here see that the product features have the correct translation.

Here, too, analogous to the workflow above, the translation is then checked, and, if necessary, changes are suggested. And if everything fits, the translated product is released and can then be output accordingly. That's it. That was a relatively straightforward workflow for product information enrichment and production. At first glance, this is often not very clear in the company or not very clearly structured. Because not only one or two employees are involved in the processing, we also have various tasks, creating products, some things are also found here again, from just. They are creating products, enriching marketing information, translating. But there might be a graphics department that must upload the images. And this department is not responsible for ensuring that the photos end up on the right products. Then we might have another task. This involves assigning or referencing the images accordingly.

We provide for the channel-specific attributes that are really in every output channel, or on a website, or if we sell via Amazon, Google Shopping, if we issue cards via Trade Bite, that all attributes are also available for these channels, as they are required. That for specific exchange formats, all information is also available. Then, an important point: Here, we still have quality checks and regional requirements, which I still include here. So that, if necessary, regional differences also flow into the product data enrichment. So. And all these tasks are not, as I said, done by one person. There are various departments involved. So, from product management to marketing, to a graphics department, to an external translation agency, to a quality assurance department, or even the legal department, which then looks to see whether we are being relatively safe for this brand that we are selling, be it chemical products, pharmaceutical products, in other words, everything that is subject to some specific regulation and is subject to that, and that we are not giving out any false information to the outside world. And from this, it can be brought down to this one denominator quite well to make this structured workflow. These tasks define certain work statuses that are to be achieved for each job. And these work statuses must then always be passed on accordingly from one responsible person to the next. And all this along a predefined processing chain.

Everything is predefined as far as possible so individual links in the chain cannot break out of this workflow, and then data that does not conform to the law is passed on to the outside in the wrong place, is somehow distributed to customers. This means that the information is incorrect, giving rise to the recourse as mentioned above claims. For completeness, I have included a particular PIM vocabulary here. These are three keywords that you hear again and again in the areas you have with workflows. When we deal with workflows and workflows, then there is the term workflow management in the area of PIM and also in the system of feature descriptions. This ensures that certain work statuses are then completed and passed on to the next responsible person. These responsible persons or instances are anchored in roles with certain editing rights. One person may only have read access. Other people then have to write access. And only a third instance can release a product for output to the outside world. The whole thing can then be controlled via a user definition of the user groups so that you don't always have to define the rules for individual people, but can perhaps also prevent them for entire groups. So.

If not, all cases have been adequately considered? Well, the first point, the long time-to-market, is like that: We may have one employee who has prepared the data exceptionally well up to a certain point. The product is almost ready. But now, there is still one test that needs to be done to issue the product. So, a processing status is missing. And then, it has to be ensured that this product does not simply sit on the shelf for an unnecessarily long time and cannot be output. Only if the information is placed externally can the product be found and finally purchased. Thus. That means that if necessary, the around the conversion is generated only later. I have already mentioned the problem: if we have a side path through our workflows, there is still such a side path that product information gets past the actual quality assurance into specific channels and ends up with the wrong information at the customer or in the store. Starting simply with incorrect product images, but also merely when certain features are misdescribed. If size specifications are faulty or voltage specifications are wrong, then the product is purchased and installed, and, if necessary, the damage is caused to the customer.

The initial consequences of this are a diminished or poorer reputation. Then there are returns, which generate new costs, and then there are the claims above for recourse. So when things get dire. This uncoordinated processing of product data leads to customer dissatisfaction, unnecessary stress, and inefficient work for the employees. And to straighten this out, the PIM can be a good help if we can create the workflows here.

That is, starting from the left, we have faster workflows. As a result, efficiency goes up, of course. The process costs decrease. We have legal certainty that quality assurance is adhered to, and only legally compliant data is passed on to the outside world. Or, if incorrect information is available, if product specifications have changed and need to be changed quickly, this can be done promptly in the corresponding workflow. The data can be corrected promptly. They can then be distributed to all channels. Into this issue, that also brings us to the paper aeroplane icon: We have with the PIM and the corresponding workflows also the possibility to obtain the data quite unerringly after the update into the related channels, by using the corresponding, released data. And as a result, not only the employee in question is satisfied, but also the customer, who then really gets what he expects and is satisfied and happy with the purchasing process and can then become multiple buyers again. To design the workflows accordingly, we at diva-e see a dedicated workflow workshop in the PIM project, divided into a scoping and concept phase and a corresponding implementation area. This is also framed here on the left-hand side, where we then really enter into conversation with the selected roles in the company. That we look at Who is involved in the processing of product information? What editing rights, what reading rights must be assigned? Who has access to the information and when, so this is really defined and can then be implemented accordingly? Once this rights and roles concept and the workflows have been implemented, we usually meet with individual representatives of the various user groups and put them through their paces again to see whether the workflows that have been designed and set up are secure and sustainable. And only then does the corresponding live-run take place in the project, with further development potential. And then, we look at where additional workflows can be added and optimized. So. In this workflow workshop, as I said, our goal is to take the customer by the hand, get to know him, and see how the data processing is running. That is, how the actual process is and how the target process can best look. Where does the data come from, where does it go? What else needs to be done in between? Who is responsible for collecting the data? Who has which enrichment tasks, i.e. marketing information, technical information? Who is then allowed to release specific information packages? Who ensures that a product can go to the outside world? And who then releases the product to the outside world? In addition to the internal players, external players also play a role, such as the translation above agencies, which can then be involved so that we can look at: Who needs to be involved now, who needs to have access to what? Once the data output has been designed and we have clarified this workflow, we must also ensure that if a task is not completed on time or wants to break out of this workflow, what escalation mechanisms are then in place. This starts with an application, a message, and the marking of a product. There is still work to be done that specific tasks still pop up and appear in the list. That the employee still has something to do. And that if that still doesn't happen, the next escalation level is reached and, if necessary, an e-mail is sent to the next higher-level person. Or that the task can be assigned to someone else if necessary. So that as minor time loss as possible here and simply efficiency can be optimized. The last point that I see here: If you look at the workflow, we imagine processing. Then there's someone who checks it and then releases it. But it's not always entirely linear, and it can also be that specific tasks are processed in parallel. That work is done parallel on texts, images and attributes, and that this data is then merged at some point. But before that, things run in parallel and not necessarily one after the other. Exactly. And then it's simply a matter of determining that and finding out in an individual conversation and finding the optimal way like this. Oliver will now show us how the whole thing can look in practice with the workflows and optimize the workflows and possibilities in the PIM. Then I would like to hand over the floor and the presentation to you.

Oliver Kleinjans: Yes, Markus. Thank you very much for these insights into your approach and for your deep understanding of what has to be considered when designing such workflows within the PIM system. I will now try to bring it to a practical level in an example. Short starting situation, so that it is also comprehensible for every participant: We have set up an e-commerce retailer as an example use case, which has specialized, among other things, in the sale of TV sets. In preparation for a new product launch, the product manager must identify appropriate products, which are then enriched by the marketing team to the degree of completeness defined by the manager in product data. And finally, the manager has to approve these enrichments to publish the new TV sets. Two user accounts are used for this in the use case.

Meanwhile, I'll switch again to the marketing colleague to start the data enrichment and the release request once the product manager. But based on the initial situation, we'll start with our Product Manager system. This logs into our akeneo PIM system. It wasn't planned that way. Unfortunately, that happens now and then when you want to do something like this. We are currently in the login screen of our PIM system, within my account, me as the product manager. In the products area, I, as the responsible product manager, now want to select my TVs to prepare them for publication.

I have as a filter option for this, for example, my category tree, which I can use to jump directly to my category of TV sets. This reduces my product selection here in our right display area. But can, of course, also use all the standardized filters, such as a completeness filter, creation date, edit date. But I still can filter on any attribute that I have in my PIM system to target and identify the products that I now want to have enriched for my publication. Bring them to the level of completeness that will allow me to publish in the first place. As we can see here, we have been able to identify two TV sets via our selection in the filter area that is not yet ready to be published in terms of their completeness. As a product manager, the entirety defined in advance is still at 28 per cent for both products. And so, I don't want to give product data to the outside world. I want to provide complete product data to the outside world. Markus has just explained it to us very well: A lot is riding on it! I want to give the customer all the information he needs to make a purchase decision.

Here, too, we see: The quality is only at a grade E. This is the worst grade in our system. This means: As the person responsible for product data, I don't want to find my product like this in the front end of my store. To optimize these products now, I create a so-called project. Within this project, the product selection I made is taken over. For this, I go, complete the project, give the project a project name. Of course, this task should also have an end date. And I can provide the collaborating colleagues within the project here still a description with to the hand, what should take place here. I saved the project, in the background, in the system background. The project has now been set up so far. Through the assignment of user rights, the system automatically knows which users must now participate in this project. It notifies these users internally that a new project exists and, of course, passes on all the information that I, as the product manager, have just entered in the login mask. To do this, I now change the account to the person responsible for data maintenance. In our example, this is Julia. Julia logs on to the system. And already in her dashboard, she sees: There is a new project. But also has a notification in the system itself, with this red dot indicating that there is something new, unread. That new products need to be enriched for the "New TVs" project. And that the due date is 7/16. In addition, she has all the necessary information to do that in her login dashboard. She sees the name of the project. If she is actively working on multiple projects, she could now switch between her projects. She considers the current processing status. Zero products have zero maintenance. So, there is already some level of care everywhere. Therefore, two products are within processing, but still, zero products are finished. The whole project is in our output channel, e-commerce, as mentioned in the beginning in the situation description. We have an online store, among others, for TV sets. The project's due date is 7/16 Plus, still my description, respectively the description of the product manager, what exactly should be done here. Like Julia, I now start my day. I recognize my task and jump from here directly into the project. I see the products that need to be processed and could, of course, now work through many products very quickly here using mass processing or sequential processing tools. However, to demonstrate the use case to the smallest detail, we would now only start a single product processing here. So I select my product and end up in the so-called product editing form of the individual product. What do we see here now? The product has completeness of 28 per cent. Five attributes are missing. And these required attributes still need to be maintained for the product to reach the entirety for a publication of one hundred per cent. They are each marked by an indicator within the listing of features. This means that I can immediately see which information still needs to be enriched to achieve completeness. But I see the entire spectrum of existing data that is already there and is already present on the product. Of course, in reality, such a list of product information is much longer. Therefore, it might be more time-consuming to pick out the corresponding information here. Accordingly, quite a simple remedy: Via a simple click, I can filter here to precisely the product information that now still needs to be enriched. First, we would have an attribute colour that needs to be improved. From a selection list, I now take my value at this point. The screen diagonal must still be enriched. Of course, I maintain this directly. Whether the device is WLAN-capable. I still have to assign a product name. A bit more about that in a moment: note that a rule can update this attribute.

Simple data maintenance tasks, such as linking existing product information to a name, can be done automatically by the system. I'm doing it manually here now as a step. So you can see how it works in practice if you don't use this rule. But with the note on it: it can also be done by an automatism by the system itself. But as we see here now, Julia can indeed save her editing status because, we see down here, she's in a working copy. Maybe she wants to keep now in between because she wants to get a coffee. No problem at all. By saving this, we no longer have five missing attributes, but only one, the main product image. Here the system gives me directly an appropriate hint where I can maintain this information. Namely, I have to go to the assets area in this secondary navigation here. And I can now, in the attribute of the main product image, select my corresponding asset from my asset gallery, assign it to the product, save it again. Now see that I have achieved completeness of one hundred per cent with this editing version of the product. However, like Julia, I do not have the right to create a final version directly. There has to be another review by my superior product manager. That's why I am now sending this working version for approval. I can now give my product manager another comment here and send what I, like Julia, have done at this moment as product data enrichment to my superior. I didn't have to select the supervisor because the system knows manually, just like when I assigned the project, based on user rights: Who is the supervisor, and who now only has the rights within the system to accept this proposal? So for that, I will change the user account again, back to the product manager's user account. This one now has a notification that Julia has submitted a proposal for a product from data maintenance, with the following note. He now starts to edit the submission and sees that for the subsequent development with the status "Waiting for approval", a proposal has been submitted today by his colleague Julia. This proposal is based on values within the locale German for the name. Previously, no deal was available. Julia has maintained a weight. Nothing was supported before for the selection attribute yes, no, WLAN.

Julia has maintained a corresponding value. So now, for each change made by Julia, he is given a dedicated option here to accept these changes, to the attribute, or to reject these changes, because he perhaps hopes for better enrichment there or because the information is wrong at this point. To simplify the whole thing and not have to approve each attribute with the new value individually, he sees here at the first check that everything fits so far. He can authorise all at once. He gives his colleague another piece of information.

Proposals are now processed so far. If he now jumps back to the product area, he sees that for the product assigned to the project, the completeness is now at one hundred per cent, and a higher quality level has been achieved here. He receives the information accordingly if he gets an overview of his current project. He can now go to the last instance and change the product's status from deactivated to activated. And thus, the outcome would be now ultimately released for the representation in the own Shop portal, to be able to communicate here the best possible product data stock, its customer. That was it for the demo! I hope it was exciting and instructive. As I said, this was a small use case. They can be much more complex. Markus has already thrown quite a bit of information over the fence on this. But, due to time constraints, we decided on a simple but inclusive release workflow. With that, we would be so far, though, and I would give back to our moderation, the Julia.