Content Recommendation Widget Update
The RCW or content recommendation widget is a feature that our community members have been able to engage with in the last month. It’s designed to help you discover interesting content. The selection of articles is done algorithmically based on a variety of signals around content similarity, popularity, trending topics, user interests, etc. In the initial (current) phase, we choose articles from within the same domain, but in the future, we plan to show recommendations from across HubPages and Maven. Our team of data scientists and software engineers carefully measure the impact the RCW causes, and we’ve seen a consistent increase in user engagement and revenue across the entire population of users; some sites are doing significantly well!
We recognize that there are many opportunities to improve RCW behavior. We are constantly experimenting and we hope to allow our community members to have more control over specific preferences for the widget (e.g. more tightly related content to current articles and categories that you are interested in or broader recommendations, selecting what type of content would never appear, etc.) We are open to suggestions and would love to hear more from you about how we can improve this experience.Samatha, I've never heard of the widget. Is it new?
I have to agree with everyone else. I assume that my readers come to my articles because they are looking for gardening information rather than information on pets or recipes or Sports Illustrated. Please keep the recommendations they see confined to the niche site that the readers are currently on.
OldRose, every writer on a niche site will be happier and welcome the idea, should it happen like that. Thanks.
I hope that that links will stay on topic. People who visit my pages are only interested in felting.
I think that if you switch this to showing selections from other HP sites, and from Maven, you will be shooting yourselves in the foot. At least now this is related, and helpful to readers. If you start showing health sites and toughnickel (and others) on Pethelpful, for example, I think that Google will eventually see it as a type of spamming and penalize the HP sites for these actions.
If this has been a successful experiment, leave it alone. There is no need to keep trying until you destroy page ranking.This is a terrible idea for all the reasons Dr Mark mentioned. Hubpages learned this lesson the hard way years ago. Why go backwards?
Please keep the recommendations specific to the site the article is posted on.
I also don't understand the point of letting writers control what they see in the widget. Am I understanding correctly? What matters is what the readers see. If you mean we'd have control over which articles we recommend to readers then that sounds like an excellent idea.I agree with both DrMark and Eric for the same reasons they already mentioned. If this experiment worked so well by keeping recommended content within each niche site, why change that now?
Hubpages decision-makers had a good reason for creating the niche sites in the first place. I think we should stay with that endeavor and not suddenly make each domain lose their individuality by making everything link to everything else.The challenge is getting clear to me. I agreed with Dr. Mark and the rest- Erik and Glenstok. It is better to let the sleeping dog sleep on, unless it needs a vet.
Constructive feedback is welcome, so thank you all for participating in this forum. The intention of refining this widget is to help readers find interesting content—and to clarify, the widget is not the same as an ad. You might see us occasionally boost and test content. We assess the impact of each change made to the widget on revenue, sessions, and pageviews. What we’re hearing is that you like the idea of initially showing recommendations across tightly related cross-site clusters. We think that’s a good idea.
What do you mean by "cross-site clusters?"
I think most of us are saying we like the idea of only showing related recommendations from the same niche site the article is posted on.
Anything that keeps readers on the same site = good idea.
Anything intended to shuttle traffic from one site to another = very bad idea.By that I mean closely-related Network Sites. As I said, this feature will continue to evolve. We will be testing and assessing as we go.
Will other Network Sites also be sending traffic to closely related HubPages niche sites?
Sorry, I'm not sure what the question is. Are you asking if each niche site will send traffic to other niche sites?
No. When you said "Network Sites" I thought you meant other sites in the Maven network, outside of HubPages.
So, my question was: If articles on HP will help readers discover content on Maven sites through this Content Recommendation Widget, will Maven sites also help readers discover content on HP niche sites via the same widget on their sites?
But maybe that's not what you meant. By "network" did you mean only HP niche sites?
If so, that actually makes me a little more confused. Because any closely related content from HubPages should all be on the same niche site. So, why not just have the niche sites only recommend their own content, which is what most of us are suggesting?
I guess I have birding articles on both dengarden and Owlcation. But, for the most part, closely related HP content is always on the same niche site.I meant niche sites. Some articles are related even if they're on different sites. Articles on FeltMagnet and HobbyLark have some similarities.
We are planning to have cross-platform recommendations in the future.
The thing to keep in mind is that we test all kinds of ideas, but will only keep the ones that work
"you might see us occassionally," and certainly some who would not log in for a week or two, will be at a great disadvantage.
I agree with all that the others have said, but I also know that this is a Maven thing and ultimately HP is not going to have a say in it.
My question would be, are sports illustrated ads going to be shown on Pethelpful and Dengarden or would they at least be limited to pets and gardening? These are of course just examples.
As we know the recommendations widget is c**p on the hubpages domain because there are no real related articles, I guess? I have a strong feeling when the pool is made larger with all the maven sites, etc. the recommendations we see are going to be like the recommendations we see on the hubpages.com domain right now. Is this something you think the team would look to prevent?
Or are they looking for short term increases in clicks with a long term slap from Google, which we all know is the source of most of the traffic across all the HubPages domains and probably most of the Maven partnered sites.With that noted, the question is: why did HP partner with maven? The odds were now obvious. Hubpages' just threw both arms up when it should be clear that maven is the ultimate beneficiary.
The widget is a good idea. Most other content sites have them. But I agree, whatever content is suggested should be related to the content the reader is looking at. However, I think you could cross niche sites and that would be fine as long as it's staying on topic. For instance, I have pregnancy related articles on both WeHaveKids and PatientsLounge. PatientsLounge has awful traffic but the content isn't necessarily all bad. If you were to suggest related articles on PatientsLounge to WeHaveKids readers wouldn't that be a good thing? Using high traffic niches to drive traffic to lower performing niches seems like it would be good.
I understand your point, but the problem is that Google lowers the ranking when it finds outgoing links to a lower-ranking domain.
- Will Apseposted 4 years ago
0I'm assuming that this is a variant of Liftigniter. Not much about it online but this page has some info:
https://cloud.google.com/customers/liftigniter
One quote:
After a period of development on another cloud platform, LiftIgniter entered the Google Machine Learning Competition put on by Google Cloud, winning two of the four prizes in the contest.
Another:
Improves CTR up to 240%
Of course, what works for news sites like The Street or Sports Illustrated might not work for ordinary sites like ours.
The rewards are not that fantastic for ordinary writers either. They say "Increases customers’ total e-commerce revenues as much as 10%"
That is millions for a corporation but not so much for the likes of us.
Not that I would say no to 10 per cent.
The issue is, will it work? And how much more labour will Hubbers have to put into it.
If we just end up as a link farm for Maven and suffer in Google search, as others have suggested, any possible gains will be wiped out.
My general feeling is that Liftigniter was designed to work within a particular site with skilled personnel managing it. Using it across multiple unrelated sites with amateurs at the helm is dangerous. - Will Apseposted 4 years ago0
This is the worrying part for me:
Quote:
".... we hope to allow our community members to have more control over specific preferences for the widget (e.g. more tightly related content to current articles and categories that you are interested in or broader recommendations, selecting what type of content would never appear, etc."
If you are going to let hubbers recommend whatever they want, then at best, you are unleashing a lot of well meant but non-relevant linking and at worst you will be providing a platform for spammers. - Will Apseposted 4 years ago0
Let's hope the sites recover from the ideas that do not work.
