Make Metadata work for your business in ÃÛ¶¹ÊÓÆµ Experience Manager Assets
Learn how to make the most out of your metadata in AEM Assets by reducing the workload to tag assets and by making your assets more searchable.
Robin Bussell Senior Trainer / ÃÛ¶¹ÊÓÆµ
Please welcome to your screens Robin Bussell. Good morning, my name is Robin Bussell and I’m going to talk to you about how you can make metadata work for your business. So I’m going to talk about what metadata is, why you need it and how you can control the effect of metadata on search results. I’ll also discuss search promote and I’ll talk about different types and standards within metadata. And finally, I’ll talk about tags and ÃÛ¶¹ÊÓÆµâ€™s smart tagging system. So what exactly is metadata? Metadata can be thought of as meaning data about data. So there are common metadata properties such as the asset creation date, the last modified date, the asset size and so on. And these kind of properties are generally created automatically when the metadata is added to ÃÛ¶¹ÊÓÆµ Experience Manager or uploaded into the system. Also, there are descriptive metadata properties such as the owner of the asset, a title, description and copyright details, for example. And those kind of metadata items would be created usually manually by a user or in some cases semi-automatically by things like the ÃÛ¶¹ÊÓÆµ smart tagging system, which we’re going to be talking about later on. If an asset has no metadata or the words don’t match, the words that you use in your search don’t match the keywords that are in the metadata, then of course the asset just won’t appear in any search results and you probably won’t actually know it’s there at all. So it becomes invisible. It’s sometimes known as dark assets and it’s costing your organization in terms of the storage costs, the backups, the licensing and so on, but with absolutely no return on investment whatsoever. So clearly we need metadata. Metadata makes the assets more searchable. It helps you to remove old and unused assets, as I said, sometimes known as dark assets. It reduces the size of the stories that you need for your digital asset management. It makes the assets easier to manage because you have more information about the assets that you can use as part of the metadata and it makes the assets richer and more relevant to the users. With the search feature in AEM, you’re able to search for metadata in assets using a full text search and you can add filters then to narrow down the results and save the search filter settings so that they can be reused in the future. You can also add wildcard support using asterisk or a question mark character to improve the search results. And the search feature in AEM allows for multi-facet support and it’s highly customizable so you can actually add your own unique metadata fields and include those in the filters in order to customize the search features to suit your own organization. When searches are performed on assets in AEM, it’s possible to add promote keywords, search promote keywords. And these keywords allow the assets to be pushed further up the list of results so you can put the assets nearer the beginning of your set of search results by adding keywords which are particularly relevant to that particular asset. And if you get a lot of results and you want to filter the results, you can use search filters by opening the search filter bar on the left hand side of the screen. You can then filter on particular fields. So for example, you could filter on the type of file or whether or not the asset has been published or whether or not the asset was uploaded in the last two weeks. And there are a number of other options.
Let’s look at metadata schemas. Metadata schemas are the sets of metadata that can be attached to a particular asset and they’re viewed by going to the asset properties. So if you open the properties of an asset in ÃÛ¶¹ÊÓÆµ Experience Manager, you can look at the metadata that’s available there. And within the metadata in AEM, there are a number of different types of metadata. You can use default metadata, which is standard for all assets, image specific metadata for things like JPEG and PNG assets, and also application specific metadata for things like PDF documents. There’s metadata which is unique to videos. And then there are certain types of assets within AEM such as collections, content fragments, forms which are unique asset types within AEM which have their own specific metadata. You can access the properties page of an asset from the quick actions button if you’re in card view, or you can use the actions bar at the top of the screen in the other views. And there you can see all of the metadata for a particular asset in a dialogue which contains a number of different tabs for the different types of metadata that are available. There are various different ways of embedding metadata in files and a selection of different encoding standards that are supported by ÃÛ¶¹ÊÓÆµ Experience Manager. For example, XMP, which stands for Extensible Metadata Platform, is an open standard that’s used by ÃÛ¶¹ÊÓÆµ Experience Manager assets for all metadata management. And the standard offers universal metadata encoding that can be embedded into all file formats. And this makes your metadata exchangeable between AEM and other ÃÛ¶¹ÊÓÆµ products and other products from other organisations as well. ID3 is a particular set of metadata which is useful for MP3 file format, so it’s useful for audio and video files. EXIF is short for Exchangeable Image File Format and that’s a very popular metadata format for digital photography. Generally, most modern digital cameras will generate that kind of metadata and that can be imported and used in AEM and other ÃÛ¶¹ÊÓÆµ products. Although there are one or two file formats that don’t support EXIF. And then there are other types of metadata, for example, for legacy applications and for legacy systems such as Word, PowerPoint and Excel where they have their own styles of metadata. Within the metadata, there are a number of different asset property groups. The first tab in the metadata dialogue is called Basic and that shows common descriptive metadata and references for the assets. So general information like the title, the description and so on. The second tab is called Advanced and that’s for Digital Rights Management. So it contains licensing information and other details. It also contains a promotion keyword field which we’ll talk about in a moment and see in the demonstration. Then we have the IPTC and the IPTC extension tabs and these contain metadata which is associated with that particular standard that I mentioned a moment ago. Then there’s a field for camera data. So that contains information about the camera that took the picture, the shot details, the aperture, the shutter speed and so on. There’s a field for insights which gathers information from ÃÛ¶¹ÊÓÆµ Analytics and displays information about how many times the assets has been viewed or clicked through, for example. And finally, references gives information about how the asset has been referenced. If you’re using ÃÛ¶¹ÊÓÆµ Experience Manager sites, for example, you can see the references to an asset shown in the references column. I mentioned IPTC and IPTC extension data and this is data that’s been standardised by the International Press and Telecommunications Council and contains subject specific metadata for things like the contact details of the photographer, the image details, the content status, particular keywords and description details and you may have details about a model if there’s a model in the image, for example. And this is particularly useful for exchanging of information between organisations as the name implies, particularly in the press, but also by other organisations as well.
So let’s switch over now to a demonstration of how the assets can be manipulated. We’re just going to have a look at adding some details to these assets. So as you can see, I’ve got here a folder which contains a number of assets which contain rucksacks. And so at the moment, these assets don’t have any metadata or very little metadata in them. If I have a look at this metadata for this particular rucksack, for example, you can see that the title and description fields are empty, but there are some smart tags and these have been created automatically by the smart tagging mechanism. And so this has already been tagged with the word rucksack. If I have a look at one of the others, for example, the equipment image here, you can see that the equipment image doesn’t contain the keyword rucksack in the tags or anywhere else. And so naturally that won’t appear in search results. So if I were to perform a search, if I just go back to the top level assets folder for a moment and perform a search for the word rucksack.
Oops, helps if I can spell it.
Then the results do show the four rucksack images which are tagged with smart tags saying rucksack. So they appear in the results and there you can see an example of how smart tagging makes those assets visible. We also have a content fragment that’s come up because that also contains the word rucksack, but we don’t have the other images because they didn’t contain any metadata with that word. So now what I’m going to do is go back into the folder and improve the metadata in these assets. I’m first of all going to select the four rucksacks and add rucksack as a title here. And you notice that you can bulk edit properties of assets by using the bulk edit feature. You can select multiple assets and then you can update the information for those assets all in one operation. So let’s give them all the title of rucksack. I won’t fill in any other details at the moment. They’ll just simply the title and then save and close that information in append mode to append those changes to what’s already there on the assets. Now I might also decide that actually I want this particular picture of the equipment to be highlighted when the keyword rucksack is searched for. So what I can do is go to the advanced tab and I can add a search promote keyword here for rucksack and that will cause the rucksack in this image to be higher up the list of results, which we’ll see in a moment.
And then finally I like red rucksacks so I’m going to make the red rucksack a little bit more relevant by adding a description here. I’ll call it rucksack for hiking. So by having a title and description so the word rucksack is mentioned twice here that will push the red rucksack higher up the list of results. So now let’s just try that search once again and look at the results. And as you can see that the first result is the one where I’ve used a search promote keyword. So search promote keywords push your assets to the top of the list. The second result is the red rucksack because I added both a title and a description in the properties of that rucksack. And then the others appear because of the fact that they contain both the title rucksack and also they contain the smart tag for rucksack.
But the other asset the one where I had with a picture of some people walking with a rucksack on the back that one I didn’t add rucksack to as metadata and that of course has now become a dark asset. I can’t find it with that keyword unless I were to improve the metadata on that asset. Okay let’s go back to the slides now and I’ll stop sharing there.
And let’s continue with the next slide. So mechanisms for updating asset metadata.
I demonstrated in that short demo just there that you can update the metadata of assets in bulk by simply selecting multiple assets and then adding the metadata to those assets in one operation. And that’s a very powerful way of being able to update a lot of assets in one go. For example imagine you want to add a tag to 100 assets you don’t have to go to each one in turn and select it and at the tag you can simply select all that 100 assets and then add the tag to all of them. And if you use append mode the other tags will be left there and the ones you’ve added will be appended to those assets. Another way of updating asset metadata is to export the metadata, make a change to the metadata outside of AEM and then import the changes back into ÃÛ¶¹ÊÓÆµ Experience Manager. And when you do that you export the metadata in the form of a comma separated variable file so it can be edited with tools like Microsoft Excel for example. And the third way that you can update asset metadata is to add a metadata profile to an asset folder and that means that you can then upload assets into that folder and as they’re uploaded the metadata will be attached to the asset automatically. So that’s another powerful way of adding metadata automatically. That could be particularly useful for things like adding copyright notices onto your asset so every time an asset is loaded uploaded into AEM it’s given a copyright message for your organisation.
So the bulk update mechanism just a moment the bulk update mechanism as I mentioned a moment ago I demonstrated just now that’s where you select multiple assets and you can update the assets by adding metadata into the various fields. The asset metadata export and import feature allows you to export the metadata as I said in CSV format and make a change to it and then import it again. And when you do that the metadata export and import operation creates a metadata job. So the job is exported and imported and you can actually see the job appearing in the jobs console which is part of the assets console.
So once again let’s just switch back to my machine and demonstrate that.
And what I’m going to do here is to first of all just export the metadata for the folder of assets that I’ve currently been working on this skill builders folder. I’m going to go up a level so I can select the folder itself and then select the skill builders folder and notice that export of metadata is done on a folder basis not on an individual asset basis so you select the folder here then I can choose the option to export metadata which is on the top of the screen. And I can give the metadata export a file name. I can include subfolders optionally or not include them. I can specify for this export to happen now or later and I can select if I choose later I can select a particular date and time in the future for the export to happen. But I’m just going to do this now because I want to demonstrate it immediately. So I’m going to select the properties that are going to be exported. You can choose to export everything or you can choose selected properties. So for the sake of this demonstration I’m just going to select specific properties from these assets for their title and description. And so the title property is called DC Title. DC is the namespace for doubling core by the way and the other property I want to export is the description. Now I can click on export and the export will be carried out. It tells me that I’ll be notified on successful completion but because it’s a complete export I can click on export and the export will be carried out. So I can click on export and the export will be carried out. It tells me that I’ll be notified on successful completion but because it’s only a few assets and a few fields that would be very quick. So now in just a few seconds if I click on the inbox notification you can see there there’s a message saying success and I can click on that and select the csv download button in the top left corner and download the export. So I’m saving that to my local downloads folder. Let’s open that up in Microsoft Excel now.
And there’s the result.
And as you can see we’ve got a column called asset path which contains the path to the assets that were in my folder and I’ve got the DC title and DC description fields here. So imagine I wanted to do some changes. I wanted to make some changes to these fields. What I can do is I can edit them here in Microsoft Excel and of course that could be a lot quicker than trying to edit them individually within AN. So backpack A is actually a wine rucksack. So I’m just going to label these as wine, wine coloured. Backpack B is khaki. I should have said rucksack khaki there really. Backpack C is a green rucksack and backpack D is blue.
So what I’ve done there is just updated the title field for the four rucksacks. I could carry on and add titles and descriptions to other fields here if I wanted to and they would then also get updated. So now what I do is save that data in the same format as it was originally created.
Just replacing the original file.
Then I can go back to ÃÛ¶¹ÊÓÆµ Experience Manager, drill down into the folder that I was exporting the data from and inside the folder I can choose create metadata. Now I select my file in the downloads folder and click on import to import that metadata.
When the screen is refreshed you notice that the titles of these rucksacks have now been added. Rucksack blue, rucksack green, rucksack khaki, rucksack wine. Those have all been updated from the changes I made externally. So that’s a very powerful way of being able to export and import the metadata in order to make bulk updates. Another technique that can be used for adding metadata as I mentioned earlier is to add a metadata profile and this is a profile that can be added to a folder so that all the assets uploaded into that folder will inherit the properties, the default values for the properties that you specify in the metadata profile. So I’m going to go back to the folder and I’m going to add a profile. You can create metadata profiles in the folder called tools, assets, metadata profiles and there you can either create one from scratch or you can copy an existing metadata profile and change it to create a new one. And then once you’ve created the metadata profile you would add the metadata profile to a particular folder so the metadata profile applies to everything uploaded into that folder. So there you can see in the screenshot you can do this in two places. You can do it in the metadata profiles console or you can do it on the actual asset folder here. So once again I’m going to switch over to a quick demonstration of that and what we’ll do is switch to the tools part of ÃÛ¶¹ÊÓÆµ Experience Manager, then go to assets and metadata profiles.
And as you can see there is already a metadata profile that’s been prepared for this demonstration called ACME DRN Profile. So if I edit that you can see that we’ve pre-configured this metadata profile with some default values. We’ve given a description of sports image, I create a tool of Photoshop both on the basic tab and if I go to the advanced tab you can see that the other properties for creator, contributor, copyright, copyright owner, usage terms, these have all been filled in with default values for the organization called ACME here and the names of the people involved. So I’m just going to leave those changes. I’m not making any changes to that at the moment but what we need to do now is apply that metadata profile to a particular folder. So I can go down into tools, assets, metadata profiles and select the metadata profile and then apply it to a folder and the folder will be the one that I’ve created for this demonstration called skill builders.
Having done that if I switch back to the assets folder and go up a level you can see that skill builders now shows that it’s got the ACME DRN profile applied to it. It’s mentioned on the front of the folder. So now if I upload an asset into that folder let’s see what happens.
So I’m going to create a new file and I’m going to use the basketball asset I’ve got here. Let’s upload that asset and once the asset’s uploaded it will take a few seconds to process. As you can see it’s showing that it’s processing at the moment. If I try and look at the metadata immediately it may not have been updated yet. So if I go to the advanced tab here you can see that at the moment this is empty but let’s just wait for a few seconds and if I just refresh this window you can now see that the metadata has been added. We’ve got a description there sports image, we’ve got the creator tool Photoshop and on the advanced tab we’ve got the creator Scott Reynolds, contributor Alison Smith and so on. So that detail has been added automatically to the asset as it was uploaded and as you can see it’s another very powerful way of being able to keep good metadata on all of your assets especially when you have a set of folders with a hierarchy where assets are sorted into different types of asset.
Okay so there’s a little bit about creating metadata and adding metadata to assets. For the remaining part of this presentation I want to talk a little bit about tags and tagging. So what exactly is a tag first of all? Well it’s a piece of metadata, it’s a kind of metadata which can be assigned to an asset very much like a luggage tag is a way of labelling a piece of luggage and you can classify content quickly and easily within a website by using tags. Tags can be helped to maximise searches to find relevant related content and a tagging taxonomy can be created by a tag administrator. So a tag administrator is able to go in and specify a hierarchy of tags starting with a namespace at the root and then drilling down into more details inside the tag information.
So you can add a tag to an asset by going to the assets properties opening up the metadata window and on the basic tab you can simply select the tag from the tags window. And those tags will be coming from a taxonomy that’s created by using the tags administration console which is part of the tools in AEM. In the tags administration console you can create a hierarchy of tags and the tags can demonstrate the kind of content you’re dealing with. So in this example you can see there we’ve got a tag console which starts with a tag and then you can add a tag to the activity tag. So in this example we’ve got a tag console which starts with the namespace called weekend which is the demonstration website used for presentations and demonstrations with AEM. Under that we’ve got a tag called activity so that’s a general tag for showing all kinds of activities and inside that we have tags which are children to the activity tag by creating a taxonomy like this it means you can add a tag to an image which contains lots of different activities which is just the tag activity or you can you can add a very specific tag for a particular kind of activity like cycling, hiking, surfing and so on. And of course you could take this further and break down the tags into even more detail you could have mountain biking and road racing for example under cycling or you could break down the other tags into even more specific groups. So by creating a taxonomy like that you can label your assets with a very well-defined description of the content of the assets.
When you’re working with the tagging console you can set up a namespace in the first place and a namespace is used to categorize a group of tags so that the tags can be organized hierarchically below that namespace. Looking at the details in the tag administration console you’re able to view the properties of a tag to see what it was last edited and the number of references to it and you can also go and look at the detail of the references to see where that tag has actually been applied in ÃÛ¶¹ÊÓÆµ Experience Manager. So let’s have a look at the tag console now. I’ll share my screen again.
And I’m going to go to the tools part of AM once again.
And under tools general we can open the tagging console.
And as you can see under weekend as I already mentioned we’ve got weekend activity, season, region, customer journey. Under activity we have tags for cycling, hiking, surfing and so on. So let’s add a new tag here for basketball as a new activity.
And having added that tag you can see it appears now in the taxonomy there. I can now go back to my assets and I could select this particular asset the basketball image and I could add that tag that I’ve just created to this image. And now of course anyone who searches for basketball would find this asset because of the tag there.
Now that’s all very well that’s a very powerful way of adding tags to an asset in order to make it easier to find but adding lots of tags can become somewhat laborious. So in practice it’s very convenient to be able to make use of smart tags. Smart tags is a mechanism powered by ÃÛ¶¹ÊÓÆµ Sensei Artificial Intelligence that allows you to augment the search capabilities of AM by automatically tagging assets based on what’s in the image. Looking at the screenshot that we’ve got on the slide I’m showing at the moment you can see there that this is a picture of an office lounger in what looks like an office building and there you can see an example of the kind of tags that are created by the smart tagging mechanism. So it says chair, office, furniture, business, seat and so on and so on and they’re pretty relevant tags in most cases. Now the artificial intelligence is not 100% perfect so it does make mistakes and you might have noticed that one of the tags for this particular picture says hair which is probably not all that relevant in this particular case and there’s one or two others that you might decide are not so relevant. So it is possible to go in and fine-tune the tagging mechanism and you can train it to become better at tagging and you can also refine the results for each image as they’re generated. So smart tags can be managed. You can promote the most relevant smart tags to make them more important in search results and more likely to be used again in other similar images. You can delete irrelevant smart tags to reduce their effect on the image and you can also add your own manage tags there as well if you wish to and you can if you make a mistake you can go back and reset everything back to the default settings. So let’s go back to the image and first of all I’m just going to illustrate some of the smart tags that have been applied. So if I go into that lounger image for example I can select the image and I can click on manage tags and there you can see quite a lot of relevant smart tags chair, office, furniture etc. But some of these are not so relevant so for example it’s not an armchair let’s get rid of that one. Hair is not really relevant so let’s get rid of that one and computer I don’t know I suppose you might have computers in your office but that’s not really relevant to this particular image so let’s remove computer there as well and you could you might want to remove one or two others. Other tags you might want to promote so chair for example office could be promoted, furniture those are probably pretty relevant as tags.
Then by doing that you’re not only making the tags more relevant to your asset but you’re also training the smart tag mechanism to become better at tagging your assets. You can also run enhanced smart tagging and in order to use enhanced smart tagging you can train the smart tag mechanism and to do that you need to add a group of tags in what’s called a tag mechanism. So you can do that by using a group of tags in what’s called a tag model to the smart tag training mechanism and then you need to provide a set of assets to which those tags should apply. So you teach the smart tagging mechanism that certain types of assets can contain certain types of tags. You can then activate the the DAN smart tag assets workflow if you’re using AM 6.5 or you can use the smart tag training console and in AM as a cloud service and they will then train the smart tagging mechanism to recognize your assets. Now this is particularly useful if you have a lot of assets which are unique to your your business or your business sector. For example imagine that you’re your business or your business sector. For example imagine that you’re an organization that sells a range of products and those products have particular product names. You might want to use the product names as tags and then when there’s a picture of a product which is of that particular brand name you want to apply that particular tag to it. So you could train the tagging mechanism to recognize pictures of one product against another and by having a set of similar but different assets that relate to the tags that you’re interested in you can train the system to become more able to recognize any new assets which again contain the same kind of images.
So these are some screenshots of the smart tag training console and as you can see there you’ve got a particular console that’s available in AM as a cloud service which allows you to configure and run smart tag training. As I mentioned a moment ago if you’re using AM 6.5 you actually do this by running a workflow. Now when you’re carrying out smart tag training there are some best practices that should be brought in mind. You need to make sure that you use unique tags so you need to make sure that you’re using the best tag for the particular image. So identify a set of unique tags that best describes the objects in the images in the context of your business. The number of images should be a minimum of 10 but probably around 30 is preferable and a maximum of 50 images per tag. The images should be visually similar for a single tag model.
You should use a sufficient variety of images for the tag model so that AM can focus on the right things. So if you’re applying the same tag on visually dissimilar images then you need at least five examples of each kind of visually dissimilar image in order that the tagging mechanism can recognize the similarities between images and then apply the tag to that images with a similar appearance. You should use images which focus well on the subject with less distractions so you don’t want a lot of background noise, you don’t want a lot of people in the background that might be distracting from the main subject of the image for example. You need to add all applicable tags to an image. If you don’t apply a tag to an image then it will be considered a negative so you need to make sure that all applicable tags are applied to the image in order for it to train and learn to use the right tags. The number of tags should have at least two distinct tags and at least 10 different images for each tag and there should be no more than 50 tags.
And the number of examples you should have at least 10 images for each tag but ÃÛ¶¹ÊÓÆµ recommends 30 images.
Finally to prevent false positives and conflicts you should create a single tag model with a single visual aspect. Don’t use a common tag for two different tag models, an AM will overwrite one with another for a common root keyword. And finally the images, you can use the same images to create different tag models if required but don’t associate the same image with more than one tag in a tag model. So those are just a few details on smart tag best practices.
And that’s just about it for my presentation so just to summarize what we’ve covered we’ve talked about the fact that assets metadata are essential to manage your assets efficiently.
AM can categorize, organize and search for assets based on the metadata properties of an asset.
You can use bulk updated metadata by selecting multiple assets and updating them in one operation. You can import and export metadata in a CSV file format and use things like excel to update the metadata and you can use a metadata profile to apply default metadata to assets as they’re being uploaded into a folder. You can also create tags and taxonomies to classify the content with your in your website and you can do that manually or you can use the smart tags feature which allows you to add metadata tags automatically to images videos and text based assets which are uploaded into AM assets. And that completes my presentation so if anyone has any questions please feel free to add the questions into the chat window and I look forward to answering them.
Okay I can’t see. I have one question from Thomas. He asked if I had an image in Photoshop would the metadata be recognized by AM? Yes thanks Thomas. Yes because the metadata in Photoshop would conform to one of the standards that I was mentioning earlier the data would be recognized in AM and you would see the information being imported into AM. Any other questions? I can see no new questions.
There’s a new one. Oh perhaps my screen isn’t refreshing let me just. No that’s not.
We still have a couple of minutes if anyone has any questions.
We have another one from Thomas. Is there a direct connection possible between Photoshop and AM? Yes you can link with ÃÛ¶¹ÊÓÆµ Asset Link. You can link to Photoshop and you can actually edit assets within Photoshop that have come from ÃÛ¶¹ÊÓÆµ Experience Manager and then update upload them back into AM very seamlessly using ÃÛ¶¹ÊÓÆµ Asset Link. Thanks Robin. I think I mean if we don’t have any other question I think we can close the session. Okay and maybe start to stay online a couple more minutes just to see whether there is any further question and you can type directly on the chat. Yeah sure. Well thank you everyone for listening. Hope you found it useful. Thanks everyone.