Answers to your Quick Start wizard questions
We use a third party provider to crawl your URLs and extract content from your page to create items for your Bibblio catalogue. It will only import URLs that loaded with our Quick Start code, and only process data required to generate and display recommendations: the URL path, the text, a title, an image, a publication date and an excerpt, if available.
Our standard snippet uses your canonical URL values to assign a unique ID to your content items and ensure we don't import a page more than once. You can however still use Bibblio without them by customizing the snippet to add your own set of IDs.
Once imported, the items will be matched against the rest of your catalogue and recommendations will automatically appear where you added the second snippet.
If you wish to use your own fonts, or make any other changes, you can do this within your own stylesheet. Be sure your stylesheet loads after ours, and include !important in the parameter to guarantee you'll override our settings.
If the thumbnails are displaying a grey background, it means we couldn't find an image on your page to use for that specific content item.
When your articles do not consistently have an image we recommend you add one or several placeholder images to your stylesheet that will load instead of our grey background image. Get in touch if you'd like to do this but don't know how.
If the recommendations are consistently pulling the wrong title/image/author/date attribute, let us know so we can correct it.
It is currently not possible to change the thumbnail image for an article or edit the recommendation title on an ad hoc basis. We are working on updating your Bibblio dashboard so you can make edits to the content we have imported. This will include updating thumbnail images, dates, titles, and more so watch this space!
If you are using a standard Quick Start code on your content pages you can add a different snippet to your homepage showcasing popular recommendations instead of relevant ones. Use the Quick Start code generator in your Bibblio account to create a new snippet and make the following edits:
-add data-recommendation-type="popular" instead
This will only work if you've had a standard Quick Start code on your site for 24h. After all, we need to gather some click data first to know what's popular...
Bibblio is compatible with AMP using an amp-iframe. You will need to have access to your AMP template and use a different snippet available on our GitHub page. Just make the following changes to the AMP snippet:
This will ensure we automatically import your articles and retrieve recommendations.
With our current version, you can have the title, excerpt, published date and author displayed. You can also choose to show an information icon at the bottom of your module with a link to help your users understand what these recommendations are.
Yes, you can separate your content into different folders or catalogues. This is helpful if you have different types of content that shouldn’t appear in the same widget, like videos and articles, or French and Italian versions of the same article. Just add data-auto-ingestion-custom-catalogue-id="YOUR_CATALOGUE_NAME" to the import snippet in 4.a in order to automatically create and import your content to a catalogue. A page that has been imported to a catalogue will by default only show recommendations from that same catalogue.
You can choose to display Bibblio on all pages but exclude some content from appearing in the recommendations. For example, you might want to only recommend articles both on your articles and on your product pages.
In that case, you'll need to create two import snippets, the first one with the property data-auto-ingestion-custom-catalogue-id="Articles" which you can add to your articles template. For the second one change the property to data-auto-ingestion-custom-catalogue-id="Product" and also add a second property data-custom-catalogue-ids="Articles" to the snippet. This can be added to your Product template and will tell Bibblio to import a product page to the "Product" catalogue but fetch recommendations from the "Articles" catalogue.
By default, your recommendations are influenced not only by how semantically similar they are, but also by how popular they are with your audience. This optimized algorithm learns which recommendations will generate the most clicks for a piece of content by exploring the content space and making predictions based on interaction data.
You can change to a non-optimised algorithm to see more semantically-related items only. We will still collect data for our machine-learning model via this method, so that you can revert to the optimized algorithm once it has enough data to make strong predictions.
This may also be useful if you have more than one Bibblio module on a page - they will show different recommendations, so your readers will be offered more variety.
For more information on how these algorithms work, read how we built it.
Our modules won't detect other Bibblio modules on the page, so any additional widgets will show the same recommendations by default.
You can customize the snippets to influence what content gets shown, as described above. For example you can choose to have one of the widgets display semantically similar content only by adding the property data-recommendation-type="related" to the snippet.
You can also choose to use our catalogue functions to display different types of content depending on the module, for example sponsored vs editorial content.
You can prioritize recency so that posts published within a chosen time span are more likely to appear as recommendations. (Older articles may still appear, but only if they're extremely relevant.) Factors such as frequency of publication, total volume of articles and value of your historic content can help you determine which setting will work best for you, or whether you need to influence your recommendations with recency at all.
For example, if your content is evergreen (e.g. recipes, DIY tips, arts and crafts etc.), we recommend switching the setting off completely. News-oriented sites with fewer than 5 articles posted a day will be fine with a setting between 14 and 30 days, while sites with highly perishable content might want a more aggressive setting, between 1 and 7 days.
If you’re still not sure, don’t forget that nothing is set in stone - you can adjust the setting at any point in the Content section of your Bibblio account. Changes take 24 hours to come into effect, so we recommend waiting a few days to see how it impacts your site.
Yes, you can. Just add data-hidden=true to the snippet tag. Note that during the installation, a grey box will appear instead of the module with the message “Bibblio is busy indexing content”. If you want to hide this from your users, you can use that parameter. Don’t forget to remove it after 48 hours, at which point we would have generated an initial set of recommendations for your site.
We do not currently have an interface for updating your module’s design and settings, so you’ll need to generate a new snippet and re-deploy. You can find the module designer that will generate the snippet in your Bibblio dashboard under Quick Start Code.
Our JS will only execute once your page has finished loading in order to keep your site fast.
The images we use for the thumbnails are served from your site, so if you are concerned about speed test scores, make sure the first image in your posts is not too heavy and in a next-gen format.
Any account on a paid plan can opt-in to have thumbnail images hosted on our servers to increase page speed.
You can test us on a staging site as long as it isn’t password-protected and you have whitelisted it on your Bibblio account. Don’t forget to visit the URLs to allow the scraping to kick in. We need a minimum of five URLs to generate recommendations.
We recommend you create two separate Bibblio accounts for your staging and live environment to avoid articles on your live site linking to staging urls.
Once you drop the code on your site, we need to scrape your pages and generate recommendations. During the first installation it may take between ten minutes and a couple of hours, depending on your traffic and the size of your corpus. If you have more than a few thousand content items, we recommend importing your posts whilst in hidden mode.
Once the module is live and we have some click data, we will generate a set of popular recommendations across your whole corpus. Any newly published article will show these popular suggestions automatically, while the actual recommendations are being generated, so you’ll never miss out on engagement.
You can add tracking parameters to your snippet using the third step of the wizard. Under Add Tracking, click Create new tracking parameter. Add a name for the parameter and its value, then click Add to snippet. A line in your snippet will look similar to this:
Repeat this process for each parameter you would like added.