Blog
How I structure Sitecore Templates

How I structure Sitecore Templates

Templates are the building blocks for a Sitecore installation, they're amazingly flexible and with capabilities such as inheritance you can produce an elegant architecture, or alternatively a complete mess. Here's how I like to structure things:

Base Templates

First off I have a folder of base templates. These are the building blocks for all the fields that will end up in a component data source or an actual page via inheritance. By setting them up as a base template we can re-use the same field definition for things like headings, but more importantly it keeps them focused on a specific purpose rather including details of the entire eventual item.

One additional rule I have for base templates is that they should only contain one data section within them. This in turn helps keep them focused on a specific purpose.

Page Templates

Next we have pages. You guessed it, a page template is what a content editor will create an instance of when they make a page. It's responsibility is to bring together default presentation details, the set of fields on the page and insert options for the pages beneath it.

All fields are inherited from base templates and standard values are used to define the defaults for the page.

Additionally to make the CMS experience as easy as possible for the content editors an icon should be set so that a page type can be visually identifiable.

Component Templates

A component template is the equivalent of a page template but for data sources.

Like page templates all the fields are inherited from base templates and standard values are used to define the defaults for the page. An icon should also be set to make content types easily identifiable by content editors.

Folder Templates

Folder templates are often overlooked but they are an essential part of creating a decent user experience.

Folder templates are created to define the insert options for components and site settings rather than having them set on the site content tree.

Where relevant a folder template should also include itself as one of the insert options so that content editors can organise their content into sub-folders.

Parameter Templates

When a component needs some config to customise its look and feel that is not content, it can be better to use Rendering Parameters rather than a data source.

Site Setting Templates

A Site configuration template is useful to contain various global settings for a site. This could include things such as the Site Logo, Google Analytics account details etc. Settings should be segmented into logical sections that have been defined in Base Templates and then inherited.

It can also be useful to separate some settings into their own template item depending on the scenario of what the setting is for.

Group Into Folders

With each these template types created, you'll end up with a tree structure that looks something like this.

Pipelines - remember the big picture

Pipelines - remember the big picture

Sitecore pipelines are great. With them you can relatively easily add and remove functionality as you wish. Pipelines like httpRequestBegin, httpRequestProcessed and mvc.beginRequest are also really useful if you need some logic to run on a page load that shouldn't really be part of a rendering. This could be anything from login checks to updating the way 404 pages are returned. However you do need to remember the big picture of what you are changing.

Pipelines don't just effect the processes of the website your building, Sitecore uses them too. That means when you add a new processor to the httpRequestBegin pipeline, that's going to effect every request in the admin CMS too. Just checking the context item also isn't enough as some things. e.g. opening a node in a tree view, will have the context of the node you clicked on!

Adding this snippet of code to the beginning of your process should keep you safe though.

1using Sitecore.Diagnostics;
2using Sitecore.Pipelines.HttpRequest;
3using Sitecore;
4using Sitecore.SecurityModel;
5 
6namespace CustomPiplelineNamespace
7{
8    public class CustomPipeline : HttpRequestProcessor
9    {
10        public override void Process(HttpRequestArgs args)
11        {
12            //Check args isn't null
13            Assert.ArgumentNotNull(args, "args");
14            //Check we have a site
15            if (Context.Site == null)
16                return;
17            //Check the sites domain isn't one for sitecore
18            if (Context.Site.Domain != DomainManager.GetDomain("sitecore"))
19                return;
20            //Check that we're not in a redirect
21            if (args.Url.FilePathWithQueryString.ToUpperInvariant().Contains("redirected=true".ToUpperInvariant()))
22                return;
23            //Check that we're not in the page editor
24            if (Context.PageMode.IsPageEditor)
25                return;
26 
27            // DO CODE
28        }
29    }
30}
Sitecore xDB - How to get count of interactions

Sitecore xDB - How to get count of interactions

If you need to rebuild your Path Analyzer maps, Sitecore have a knowledge base article on how to do it here - https://kb.sitecore.net/articles/766858

In the FAQ section it gives some info on how long you should expect it to take

What it doesn't tell you though is how you find out how many interactions your site has to process. The data is stored in analytics collection in your Mongo DB, so go there an then run...

1db.getCollection('Interactions').find({}).count()
Data Source vs Rendering Parameters

Data Source vs Rendering Parameters

On the face of it being able to specify Rendering Parameters, or a Data Source on a control achieves a very similar goal. You want to componentise your page and not fill up your page type item with fields for specific renderings. Everything should be isolated in it's own little compartment.

Linking a rendering control to a Data Source does this by moving all your data to a completely separate item, that can be re-used over multiple pages and even swapped out to do personalization or A/B testing.

Equally, using Rendering Parameters does this just as well. Just like a Data Source, you can create a template for the fields that need to be entered, and the data is kept in the same place that a link to a Data Source would be specified.

But what should I use?

The easiest way I can describe it is by asking, are you storing content or something else?

If the answer is content then you most likely want a data source. Alternatively, if it's something else like some config for a filter, or a background colour you will most likely want a rendering parameter (but there could be exceptions).

Data Sources

Data Sources have some distinct advantages over a Rendering Parameter including:

  • Data can be edited in the Experience Editor inline in your page
  • Language versions are very easy to understand by content editors
  • Easily used for personalization and testing
  • Can be reused on as many controls as you like

These things all make a Data Source a perfect option for anything content related. Without them you loose some big features related to content and ease of use. There are times non-content is also a good option. e.g. A hero banner on a page could be used across multiple pages and as well a reusing the text, if a background colour is also customizable you would want to be re-using that too.

Rendering Parameters

Likewise Rendering Parameters also have there benefits:

  • Doesn't create an extra item to publish. You just have to publish the page your editing, not the related data source
  • Easy to access in the control properties of the item your looking at, without having to find the related data source

These features make Rendering Parameters ideal for the non-content config of a control and removes non-content fields from your item templates. The thing to be aware of though is you can't easily do A/B testing or personalization with a rendering parameter, so while it's nice for your items to only contain content, that might not be the only thing that's being tested.

How about both?

There is also nothing to stop you using a Rendering Parameter and a Data Source at the same time.

You could have a situation when a Data Source is being used to populated the content on multiple different controls. Each control may have some additional config needed, such as text colour or text size. Keeping this data in the data source might become problematic as on one rendering your text may need to be black, whereas on another it may need to be white. By moving these fields to Rendering Parameters, you keep the benefit of a shared Data Source for content, while the Rendering Parameter takes care of the presentation config.