Feature Friday: The Explore Page
At USV we invest in a lot of networks and marketplaces. These are messy places where anyone can post most anything and it can be hard to find the good stuff. Enter the explore page. Using data, analytics, social signals, and often some human element, the explore pages are incredible places to discover amazing stuff. Here are some of my favorites:
If you are building a service that empowers people to post all sorts of things, I encourage you to create an explore page where your users can make sense of it all.
yup. true also on mobile. which is what we did from day one on Appsfire, pushing the button even further with buddy alerts which is a great quality signal for discovery
explore pages are even more powerful on mobilei was so tempted to screen shot tumblr’s mobile exploreit is amazing
Here’s mine:#tattoos #prose #sailing
hope you’re near power 🙂
Ha! It’s a thirsty phone, that’s for sure.
It would be interesting to do a screen shot of all the above but in mobile. The explore/discovery process is powerful on mobile because of all the natural breaks that occur throughout the day while on mobile.
This post touches on many of the items that kept me from sleep last night.The connection between the single unit of behavioral value, how and if it gets shared, and the location or collective view of the network it lives in is the philosopher’s stone of successful marketplaces and communities.
That’s a powerfully worded statement. That intersection is success, but I don’t think the user thinks of it as an intersection because you don’t stop at it. You keep going.
i didn’t sleep too well last night either. but i think it was a slight reaction to the flu shot i got yesterday.
For me it might also be that Vouvray I was sipping.
“Vouvray”.Now that’s a great way to test someone’s domain knowledge. Use a word and see if they can figure out what you mean. See if you had said “Thudnerbird” (sic) even I know enough about wine to know what you meant. But “Vouvray”, I had to google that. And if it was spelled incorrectly it wouldn’t come up in google. (Actually I just tried a few variations of “vouvray” and it does. But “Thunderbird” doesn’t even come up unless you specifically say “Thunderbird Wine”.) I know Arnold and you added “sipping” but what if I didn’t know you and you just said “For me it might have been that Vouvay I had”.A good test for CSR’s is if they can understand when a customer typically screws up a concept or word. It becomes an entire puzzle to unravel what someone means when they transpose words, letters and everything else. A big time waste (and often very funny) if you don’t know enough to parse what is being asked.
If you are into wine and live in NYC, you know bubbles.10% of the wines tasted at the major wine shops were bubbles in the last 4 months, only 50% were champagnes.And you can get a completely natural Vouvray, non dosage ( that’s geeky) for under $25.Lovely way to start an evening meal.
“you know bubbles””wines tasted at the major wine shops were bubbles”I had no idea what you meant by that. From what I can tell you mean this:http://en.wikipedia.org/wik…(Which is exactly the point I was making.)The lesson for this (in marketing) is to make sure your message is targeted at newbies if in fact you are selling to newbies. Otherwise people’s brain pauses in confusion and can’t get the rest of the messages to them. (Remember from school when you read a complicated subject that you knew nothing about and had to stop and couldn’t get the flow of the presentation because so many concepts lacked definition?) The only around this is to give the message to people who are on the level of your target to see if they understand what you mean. If you know to much it can work against you. Unless of course (as only one example) you are a restaurant or a high end product and purposely want people to be confused or want to create some cache in your product.
In my world, in the world of those that read what I write abt wine, in the world of my wine community, if bubbles is not a image, then you are simply not a target community member, reader, customer.Assumptions are the core of messaging. Inclusiveness with parameters is what creates meaning. Talking to everyone is a hopeless strategy.Or so are my guidelines ;)BTW–you are the exception in this case I believe.
Agree. And I would rarely tell someone (on a personal level at least) something that should be interpreted as not doing something right w/o knowing exactly what they are doing and why.”then you are simply not a target community member”(Agree I would totally ruin things)To wit:”The lesson for this (in marketing) is to make sure your message is targeted at newbies if in fact you are selling to newbies.”And you’re not. So this was a general observation and not targeted to what you are doing in particular and most importantly only a way to spout off about something that popped into my head. Which is the value of what I get off of AVC.
No offense meant…very sorry!Good news is that there are two really exceptional, free wine tastings, hosted by exceptionally knowledgeable friends starting in 30 minutes within 5 minutes of me that will mollify this week!And on two obscure (and wonderful) regions–the Canary Islands and the Jura. I can go deep and geeky about these and will do just that!
No substitute for hand washing. Unfortunately getting food handlers (yes i am talking to you Starbucks, as well as 3rd wave coffee shops) to wash their hands after handling money is an impossibility.
Yup. I hate that too. Another one I dislike is when they put their hand on top of the lid & where you’re supposed to drink from.
This is a billion dollar heath issue. Hey HHS, how about a PSA on hand washing? Or better yet why doesnt haward shultz et al. make it part of the pouring protocol. Seems like a no brainer! P.S. Watch how often barristers put their hands on the lip of the cup when putting it into the coffee sleeve. Speak out when you see it. (with respect of course)
Exactly. Oh I do lecture them on that.
Woah, thanks Professor Waldstein. Deep. Gotta read that a few times more.
seduced by words may be the end analysis here 😉
Arnold, would you mind expanding on this?
@awaldstein:disqus Where do you see taste mapping entering into this?I see a sort of venn diagram of showcase+curate+explore+taste map forming which I would love to see further defined.
I’ll jump in later Alex…heads down for the remainder of the day.
Discovery is powerful, but it is getting fragmented too. That’s ok if it’s “functional discovery”, ie like the examples above where they are clear unique destinations.But where discovery has failed us and is still an “explore headache” is in discovering News & Content. Of course there is Twitter, but it’s not enough, so you have to go to multiple other places like Prismatic, Techmene, Zite, FlipBoard, Engagio (for conversations discovery), – and most of this discovery is serendipitous although with a touch of social graph-based or interest-based personalization.Question is: Do users want more personalized Exploration or totally serendipitous? Or a mix of the two?
I like to mix it up a bit. It’s weird to see the same content across multiple channels…show me the stuff I missed. I think news.me / digg is on that.
People want to find what they want without looking for it at first glance. Translated–its your job to make choices for people. That taste connection is what brand is about.Have this community surfaced. Have it chosen by editors. At its core I think it’s that simple.
I think there is a differentiation between “discovery” and “exploring”. Exploring suggests digging through that which is identified as important or favorable, or at a minimum, a data point from others (curated). Discovery heads in the direction of serendipity, or as Arnold suggested, finding what you want without looking for it.I realize you recognize the difference, but – to answer your question – I just don’t know what I want all the time so I prefer to discover. Exploring may not get me to my destination, it has it’s limits.
that is how I feel about discovery versus explore. If I am in discovery mode then I am open to many topics and some of these topics might lead to further personalized exploration.
It’s a tough distinction as they might be blurred sometimes, but I understand your differentiation.
somewhere in between. It is why restaurant reviews remain popular, even though you also go to that weird hole in the wall.
Explore/showcase/featured/top-picks/hot today/staff favorites etc are all synonyms.They are all pages with units of curated content which help the network/marketplace put it’s best foot forward.An inoffensive magical way of showcasing the best they have to offer and aiding the cream rise to the top without contravening the overt democratic principles of a marketplace or open network.From a branding perspective, ‘Explore’ does have a more dynamic/exciting and self-guided vibe to it than ‘staff-picks’ but they are effectively the same thing.
But in many cases it’s automated based on algorithms or personalized settings, so it’s not necessarily staff picked. A newspaper’s articles is staff picked.
ok. so replace staff-picked with featured or showcased.I think a lot of the times these top-level curated pages are not driven algorithmically but by team members who know their audience and understand exactly what content floats their boat. Claiming algorithms is their get-out-of-jail card.
the google defense
I usually design them around algorithms with override.
Can you speak to the algorithm? Is this just a basic multiple linear regression?
Yup…simple stuff.Doing two now, one with a client one for myself on a project.For mine, I’ll be pulling events from three categories two a home page of sorts display. Their pull only ones with photos not from the supplied sets, priorities given to events from priority users, some stuff around different neighborhoods. Then manual CMS driven overrides.Nothing fancy…yet…like displaying wine that you do/should like.The magic is in making choices for the user as always.This answer the Q?
Also, we sometimes confuse Popular vs. Curated. There are several slices to exploration. I like Kickstarter’s Explore where you explore by Location, what your friends are funding, topics, etc..
are the explore pages unique to each user ?
I’ve always seen a mix.
can these explore pages be gamed for commercial gain?”Using data, analytics, social signals, and often some human element”could it become just another form of spam?
it is possible but not easy to do
‘everyone is good at something’ – Skillshare/ Voomly.i can imagine a whole black hat industry growing up around the explore page.the opportunity may be centered on how to stop it.
I agree that everyone is good at something. Don’t agree that everyone is good at sharing or teaching it though.Communication is what rises to the top as much or along side knowledge.
yes… and the thing that a person is good at isn’t always a good thing, for either themselves or for others.skillshare can be a voyage of self discovery. i like that.
“I agree that everyone is good at something.”Maybe in Manhattan there are a large percentage of people who are good at something. Or at Universities or among certain groups of people. Or on the Internet where people who are good at something are vocal and get involved or are self selecting. Like on AVC.com or any other higher end online experience. But that hasn’t been my experience though anywhere I’ve lived or from my observation of what I would call the “average” person.Special side note to anyone reading this comment. Watch the PBS Frontline on Michelle Rhee and the Washington school system. Was very interesting.
What do you mean by black hat? what variables are are used to build the explore pages? I assume that each explore page is unique to the user’s profiles?
white hat/ black hat, from the search engine optimization industry.”Enter the explore page. Using data, analytics, social signals, and often some human element”.i’m making the same assumption, and the additional assumption that each category ‘seller’ would rather see their wares discovered on the page than a competitors.
Can you speak to the data analytics approaches: I assume that they display is a function of the user? What user variables have shown to be most useful?
To drive engagement, it’s in the company’s best interests to monitor and continually improve the quality of the results.Otherwise you get craigslist or yahoo answers.
yeah, i’d like to learn more about the various ways that these explore pages are generated.
why is data, analytics, social signals and human elements mean spam? seriously?
not ‘is’ spam, but can be gamed for inauthentic discovery for commercial gain.
Explore the Community tab on Disqus folks! It’s at the top of the comments thread, to the right of the Discussion tab.
Speaking of tabs, the My Disqus tab shows I have 88 unread comments. Untrue because I’ve definitely read them. But no matter how many times I click the tab, it won’t clear the unread count unless I visit each comment again.You’re not making things easy for us anal retentive types. 😉
When you click on the My Disqus tab, are you then clicking on the sub-tab for notifications? The counter should adjust down as you scroll through the replies on screen. Please give it a try and feel free to shoot me a note jim at….
Worked this time. Didn’t the last few times I tried…so I just ignored it and the number kept climbing. Thanks. 😉
I’ve actually learned quite a lot through the community tab. Really! most people here read interesting things
I’ve never gotten into it. Not sure why not.
Great to hear!
In my popup I put a message for a few days saying “send me an email and I’ll send you $5” and nobody took me up on it! Are you able to compile stats on how many people mouse over the avatars?
The behaviour will only happen infrequently — very likely when a person who doesn’t know you finds a comment of yours interesting, and wants to see more of who you are.
I’m not sure if we have that stat; may not be as insightful as other data points.As for your solicitation, are you *really* surprised no one took you up on the offer? I’m not convinced it’s a sign that no one checked out your profile.
I would expect that there would be people that would not take it seriously or be confused. Statistically (to me) it would be nice to dovetail human skeptical nature with the fact that “x” people viewed the profile. If over the course of time, say, 500 people viewed the avatar and nobody asked for $5 that would interesting..
like Christina implied I didn’t know that you can earn $7307 in four weeks on the network.
Unrelated question > Employee equity %Any thoughts on what current “market value” for employee equity for CFO, EVP Sales typesEnterprise type company, essentially pre revenue
Fred did a whole lecture on that. Not sure where he parked the video tho.
Hi Jim,Yup – I actually attended the live event @ USV last spring time…http://www.avc.com/a_vc/201…I will go back and tune in to it…my recollection was there were some broad numbers for founders, early team members, option pool in general etc.At the end Fred did a great summary/ top level cap table – it was something like20% founders40% Employees ( early employees, and option pool)40 % investors
There ya go! Implicit in this line of inquiry: Are you joining a startup?
this is a great video… worthwhile watch for any 1st time founders (or founders in general)… influenced our structure significantly
1x annual cash comp in equity value
Would that be given immediately or is there usually vesting involved?
vest over four years, first year with a cliff
That’s what I was assuming. Too much uncertainty in the situation otherwise. Thanks for clarifying.
I think Behance’s Discover page deserves an honorable mention here. They’re filtering is a best-in-class example of this type of page. Just sayin.
To mix metaphors, the “explore” tells the user where it’s good fishing in the flow that you have enabled with your peer to peer network.Still a lot of opportunity with these explore filters. I think generally speaking we are much better at enabling the flow than surfacing the good and personally relevant stuff for an user.
definitely agree. interesting things happen between the machine surfacing vs human surfacing (reddit/HN vs Discover on Twitter) – we’ve been playing with this concept for 18 months… its still fun, but def challenging
That is fascinating. Need any help?
always, tell me about yourself – btrautsc /at/ gmail dot com
i think foursquare actually answers those problems really effectively. You click on explore, and within explore you start searching. It opens up explore as less tuned to more tuned search.Probably one of the best implementations of explore I’ve seen
I am a recent convert to the Twitter Discovery feature ( I was once a hater).Keeping up on what people are ‘favoriting’ has become increasingly useful/insightful to me.
i am glad you are getting value out of it
This brings up an important message, focusing on the human element and social signals in developing the explore page. Several of the start-ups I have recently met with or have seen construct their main page or “explore page”, do NOT take this into account as much as they should. The focus needs to be placed on the ease of access and the human elements of exploring a web page. Companies, corporate sites, should begin to innovate and drive initiatives that bring about similar concepts of user experiences.
Thanks to AVC community yesterday we are FULLY aware of the desire to explore and discover Voomly experts. This “how-to” guide is icing on the cake.Really want to thank all of you for yesterday’s feedback. Extremely informative and fun to start down this road with you guys.
that was the goal of the post. i am glad it delivered on it.
The retronaut main page is essentially an explore page http://www.retronaut.com/
I think you’ve hit on a really important challenge that platforms and marketplaces are facing – how do you surface the best of the best content to the top of the user’s experience? One solution is to simply not be open to everyone, and instead curate your users early in an effort to have a bar for content quality. But, likely at some point that model won’t be feasible anymore and then the solution then becomes a design problem.Another great example to look at is Etsy. They have excellent explore pages at the top of every product category that help to show me the depth of products that they have available. Here’s an example of the “Womens” explore page: http://www.etsy.com/browse/…
i like the idea that the users curate not the operators
You showed a few established companies here but I think the “explore” section of a product is critical in start-ups, too. It acts as a great way to show people what and how you can use a new service. Those examples are often critical for success. Doing it early is hard because it means the service creators have to populate the explore area and maintain it manually as there just isn’t enough data to automate it.
I know I’m retro here but for early stage I’m a big believer in human intervention.It’s cheaper and more important it let’s you figure out what to program in.At an early stage honestly, I use people to skim for porn/spam stuff as well.
Completely agree. Great comment.
Explore is a really difficult feature to get right. Maybe one of the most difficult.There are two different classes of explore: one for anonymous users on which you have no meta-data and one for users on which you have some sort of data to personalize your recommendations. This is often logged-in vs logged-out but not always (cookies, etc).The anonymous user explore is the homepage of many websites. In this case I think the best thing to do is to make the homepage look nice without caring too much about the relevancy of the content. Etsy does a good job of this. They might choose a collection of handmade items that are all blue for example. A person won’t buy a bunch of stuff just because it’s blue but it makes the homepage look nice and then you can search or browse for what you want.If I tend to buy clothing on Etsy or steampunk gear (http://www.etsy.com/listing… ) you can show me better options on the homepage the next time I log in.With Fred’s examples, Foursquare has far and away the best explore feature in my opinion, because they know where you’re standing and they know the kind of places you like to visit (not the kind of places you say you visit, but the kind of places you actually visit). This makes the explore feature very relevant to the user and it’s both temporally and geographical relevant which is something few other services can touch.Even when you have good data, you have to take the user and how they use your service into account. Are they shopping for themselves or for a gift? Are they looking for their favorite band or trying to discover something new? What are the common use cases and how can you meet these needs.An example of a company who misses the mark for me is Netflix. In their marketing, they show happy couples and families watching Netflix together, but their recommendation algorithm assumes everyone on the account is the same person.The explore feature on Netflix is terrible for me because my wife and I watch very different kinds of movies. So Netflix wants me to watch movies with Ryan Gosling in them, which I don’t want to do, and my wife is show recommendations for Hong Kong action films. Many of my friends who use Netflix have the same complaints.I don’t know if this is a problem worth solving for Netfilx (not all problems are), but it’s an example of how sometimes user data can actually lead to worse recommendations than a simple popularity chart or genre-based browse.
So does amazon/most ad companies. The idea of multiple login/one account is a little off for these people. If you over anonymize the data, you get the problem you are talking about.The question is are you willing to be less anonymous?
From my own experience, subscription services like Netflix suffer from this the most. With Amazon, my wife and I have separate accounts because it doesn’t cost anything. With Netflix two accounts would cost twice as much for a very marginal benefit. Same with Spotify, etc.I don’t know the answer. I would need access to their users and data to make a good decision. But I hear about it often enough that I think it would be a question worth exploring through experimentation.
Interesting discussion. The difference between Netflix/Spotify and a discovery service that really truly sucks is, with Netflix and Spotify I have an answer for why the movie/song was suggested. Being completely lost provides no value whatsoever.I’ve only very recently started sharing my Spotify account with my girlfriend. Everytime I hear a Beyonce-esk song I roll my eyes, let out a rye smile and press next. Being part of a family means you expect to share things, and there is value in the thought that you are part of a family.
If you could only hear the songs I have stuck in my head.
In their marketing, they show happy couples and families watching Netflix together, but their recommendation algorithm assumes everyone on the account is the same person.Agree and with that I always thought that was actually on purpose so that people didn’t share accounts since it seemed so obvious that it was missing. Goes along with my thinking that “if something doesn’t make sense it might be because there is some piece of info missing or something about it you don’t understand”. (A distant relative of that thinking is the “fool at the table argument” which is “if you can’t find the fool at the table it’s probably you”.)In any case Netflix is making a change to that. I can’t find the story or the link but I remember seeing the other week that there was some change being made whereby you could do exactly what we all want to do – segregate history by user. (Along the lines of my prior comment I hope I didn’t dream that..)
If you find that story could you pass it along? Sounds interesting, I had no idea they were actually working on this.
Foursquare is actually the worst “explore” page for me. They tell me places to go that are a) nearby and b) similar to where I already go. That’s pretty useless to me. If it’s nearby, I would go there if it was any good.
How do you know if a place nearby is any good if you have never been there? What sort of sources do you use?
All AI driven service suffer from lack of intent knowledge. That is why no broad based graph has ever really worked.The past is only an accurate predictor of the future if you are doing now what you were doing then.Great comment Luke.
I thought it might be helpful to surface some of the thought process of a company that is still in the process of cracking the ‘Explore’ nut.We are not very good at explore/discover at Shapeways. It’s a difficult problem when you have a (theoretically) infinite set of products that can be created. Do you go the Amazon route and allow browse by category? Do you take into account the stat that only 7% of users on Amazon browse vs. using search? Do you surface the most popular of all time or the most recent?Recently, we’ve experimented with a heavily curated homepage that drives users towards one of our core categories. We’re starting to realize that a one-size-fits-all order doesn’t work very well for us, because we’re trying to please everyone. Perhaps we need to be more focused, or perhaps we are trying to solve everyone’s problem at once.What we’re lacking is any sort of context for each user that comes to our site.We’re now starting to approach this problem differently. Most users are starting at a specific product and navigating from there (as is the case with many of the sites that were surfaced above). Perhaps we can gain context from users who land on a specific product and use that to provide a great ‘Explore’ experience. We will be operating under the assumption that something about the product they landed on was relevant to them.Would love to hear any thoughts/advice/feedback.
Amazon does do recommendations. A lot of them. I find I’m inundated with aaron sorkin, bath suplies, and cookbooks on amazon on the front page (ha, it knows me well).I think getting an idea of what people buy is a starter, then segment down based on affinities (if it were me buying a lot of stuff on shapeways, it would be tea cups and jewelry if it were you, maybe not tea cups and jewelry). One of shapeways’s non-obvious problems is unlike amazon, you can’t use prime front page space to push your own thing (aka the kindle) (unless you have some object you want to push).Does this help? if it doesn’t, shoot me an email. shana dot carp at gmail
Definitely helpful Shana!! I agree, the personalized recommendation is definitely an area where we could see alot of mileage. You’re right that getting an idea of what people buy is a good starting point, but we’re seeing the primary challenge being converting a user on their first visit. Once a user has purchased, it’s much easier to drill down on what they like.On the other hand… we can do session tracking for guests and base it on visits via collaborative filtering (people who viewed this also viewed…) and also can use some of our light social actions (favorite, comment, etc) as data points.
yeah,networks can be messy. Explore and curation seem to offer a solution. Just been involved in the launch of a new social marketplace – eSeekerz.com – its great and is attracting a lof of Mandarin speakers, – but messy is a good way of describing the rush by early ‘likers’ to post anything and everything
One of my favorite examples: http://www.buildzoom.com/ga…
I was just thinking about this yesterday when I mentioned a Discover page to @andyswan:disqus for Voomly. When I was writing the comment, I specifically had Kickstarter in mind. One of the best out there. Another favorite is Twitter’s Discover. Been using that more and more lately.
yeah, the twitter discover tab is a daily visit for me, usually at the start of the day
httyr rtur u u
meetup just launched a good one (imho) http://meetup.com/explorenext up (soon)… mobile discovery
Most of the screen shots seemed like were taken on a desk top vs mobile