Spread faster


Currently we have around 200 signups per week. This means 10,000 per year. I think we can and should grow faster.

This came in today:

If I can do something to spread it faster or more let me know:)

What can we let them know?


I already said that before, it would be wiser to fix some issues like fixing the search and clean the fake/dead profiles before growing big.
The obsession of quantity is also killing me, better small and working well than big and working badly.
Previous hospex had facilities to grow but then troubles to keep their members.
Troubles mostly due to non working websites and frustrations of sending unanswered requests and hosts not receiving requests…

Every summer there are youngsters willing to hh, checking websites like hitchwiki… So advertising isn’t a big deal. However nowadays hh are mostly using CS simply because it’s more efficient.
Once again using the TR map and going trough several dead profiles doesn’t motivate people to use it often.
However using the CS app is easy and fast and with appropriate filters most dead profiles are out.

Better to have 10 000 new members with a ratio of 50 % active than 100 000 with a ratio of 5 %.
Unactive members are kind of polluting the website.

I might sound harsch but perhaps it’s time to stop playing with tribes and introduce at the very least few efficient filters in the search. Right now there is none, the online in the last 6 months is a joke.

And perhaps asking directly active TR members what they really want, more tribes, more esthetical features or a better search, a cleaner user base would be good.

This forum like the BW forum, like the CS forum are very small and not representative at all of the average member.


I have hosted in a rural area, mostly HH, all of them contacted me through CS, few of them had a TR account.
It’s easier to find a rural host thanks to TR than CS but most people usually look for big cities. The TR search being so nasty for cities they get used to CS, then even in a rural area they use CS at first then perhaps TR.

Marketing wise there is nothing worse than launching a defective product, people will usually stay on their first impression. My first impression of TR was good because I joined early when it was very small and most members were kind of active.
If I was joining today I am not sure I would stay.

I am starting to think that TR is taking the same path as BW, just wanting greater numbers and never fixing serious issues.


there seems to be quite a few feedback asking about where’s the invite page is gone. even the signup is open, it would be nice to have invite possibility aside. one feature that comes in mind to add on invites is that people you invite gets automaticly added on your friends so you have straight some contact point and feel more associated.


What bothers you about the TR search for cities? Would love to understand in detail.
(I live in a cli/shell world so i’m blind to GUI quirks.)

I actually like that i can choose what parts of town someone lives in so i prefer the TR search to the CS one. (But then again i have not even used the CS app)


213275 CS hosts in Paris.
7215 CS hosts who logged in the last 6 months.
Around 400 more or less active ones.
Around 40 would could perhaps host next week.

If it was on TR and If you had to find a host for the next days and randomly clicked on any points, the probability to find an appropriate profile would be of 40 / 7215 = 0.0055 or 0.55 %.
You would have to click on average 200 dots to find someone who could host you.
And every dot means checking the profile for few seconds to sort out if the guy is active or not.
If it takes you 5 seconds for each dot 5 * 200 = 1000 seconds or something like 16 minutes just to find one potentially interesting profile.

Now if we also take into account that you would have to send several requests, maybe 10, check the profile and write personnalized requests… You could easily waste half a day and finally not find any host.

It’s an extreme example but today it’s already annoying especially on a phone to know which dot you already clicked and which one you didn’t and there are only 40 TR members in Paris.


Right, thanks for clarifying. We definitely need to improve the experience in dense areas. What could be good ways to do this?

Some ideas:
a. indicate whether a dot has already been visited
b. better filters? but what/how?
c. also show simple list of people?

Got more ideas?

I think a. is interesting because it can be implemented with relatively little effort as it hardly requires any design or UI/UX.


thanks for explaining. In practice i never had problems finding a host in minutes with TR…


Thanks for elaborating on this, Cedric! It’s really helpful to know context and details rather than negative blanket statements which usually are not very useful to us.

The map has been slowly evolving with the number of members we have on it; there was a time when we didn’t have any filters as there weren’t that many folks on the map neither. I agree that on some busier areas (such Paris) we are now getting to point when the search needs to adapt to it differently than it does now.

I’m glad that map still works fine for some less busier areas as well for less popular tribes even on busy areas.


The quality of a network is not shown by how many people the network wants to reach but by how many people want to reach the network.

I think as long as our prime motivation is a higher number of members we are not working for a network where users will have a good experience.

A transparent metric on “how active is this member” would really help. I don’t want it to be a game, whether I can find a place or not. I already trust the people on Trustroots, that is the concept of it. If I then still have to be very decisive with each person asking for a place then it seems that I don’t trust them. Ergo: Trustroots failed.
rant on
I do host people that seem to be outside of my bubble. I do host people that have different opinions than I have. I host people that I don’t like. Because that’s what this network is about, as far as I understood. About hospitality, about tolerance and about sharing.

I would like Trustroots to be a network where it is clear if a person wants to host “foreigners”, “strangers”, “the unknown”. If I don’t want to do that then why am I here? Or how can it made be transparent what people do on this platform.
rant off

It doesn’t matter for me what “tribe” one belongs to. I don’t need that to be a filter.
Or could we make a tribe “I actually host people” :wink:


Thanks, @farbenrausch, there are quite a few points in your post to unwrap, lemme know if you feel like I didn’t address some.

@guaka was making a counter-argument to a common belief that we keep hearing that Trustroots shouldn’t grow bigger (and in fact should try to stay very small). The common reasoning is basically that a bigger network equals worse experience. We believe that “member/network quality” and size of the network don’t correlate, especially at the stage where we are (tiny) at the pace we’re growing (extremely slow). There might be a point where network quality over too fast growth becomes an issue we need to address more aggressively, but we’re far from that. :slight_smile:

Just to clarify that this is a false assumption; we (the board and folks working on TR) do follow other metrics as well and in fact, value them much higher than just total number of signups. They’re much more actionable. Member count is a “nice to have” vanity metric that folks (and media) toss around but it doesn’t have a ton of real value.

A transparent metric on “how active is this member” would really help

Do you mean that something like this could be in the profile? We do have “last seen”, “replies within” and “reply %” in user profiles. Would you like to see further statistics or these explained more?

As statistics to follow, on the other hand, there are 100 ways to calculate this. We have a few that are quite actionable:

  • Simple member retention % (“how long ago % of members was seen last time at all”) that we started to record only quite recently.
  • % of new message threads that get a reply, divided to different time buckets
  • How long in average folks need to wait for replies to messages
  • How many messages there are being sent in general

I’m all ears for ideas what else we could follow.

I do host people that seem to be outside of my bubble.

That’s great that you do. :slight_smile: Trustroots’ goal is to really try to help you find folks that are very similar to you. There are other hospitality networks out there where the “totally random and different folks than I am” -connections might be easier.

I think we could sharpen our message around that promise but I was hoping the strategy how we’ve been spreading, tribes and other small features around the site would’ve kinda given it away. :wink:

How we can improve the site to help highlight those similarities and also help folks to trust each other more is at the very core of Trustroots and I could go on talking about this for days but I don’t want to hijack this thread for that. :slight_smile:

I would like Trustroots to be a network where it is clear if a person wants to host “foreigners”, “strangers”, “the unknown”.

That’s an interesting idea! It’s probably still more effective to focus on one segment and try to make the site work very well for them, rather than optimize it for “everyone”. Good food for thought, tho!


This is just a conclusion - from a follows b. It is good to hear that it is not the prime objective because out of some reason I got that impression.

Whoooahhhh I just got nerd-sniped :smiley: Amazing work!

  • Reply % only helps when I know the absolute count of threads, it could be 1/1 and thus 100% which does actually not contain real information.
  • Last seen is quite a wrong information because all messages are put through via email. So I don’t need to access the website to read the message and can be absent for 10 months but still read all messages in the minute they are pushed for example to my smartphone.
  • “replies within” is one I like but it’s the same like the first point that it takes several threads until it actually contains practical information.

% of new message threads that get a reply, divided to different time buckets

That sounds nice! I could see that a member was active a lot in the first year, but in the second year it decreased and then stopped.

Maybe “amount of unanswered requests/messages” but that would make it harder for people that do get much more requests (possibly women, is there a way to analyze that?).

If people tend to use the Yes/No/Message reply function honestly we could use that as a metric but it is very easy to influence in by just being dishonest, only clicking “Message” instead of “No” or actually replying “Yes” with harsh conditions so that the request gets canceled.

Still, encouraging trust does most probably not work if I only see the ways to fuck a system :smiley:

This seems to not fit here but I wanted to answer here because some of it does fit. Should I push it to the “Thread about tribes”-Topic?

This is totally not visible from my point of view. Actually it makes Trustroot not foster trust but creates - explicitly - segregation. In my opinion it emphasizes distrust much more than trust. How would we generate trust between the tribes?
I mean…

What is your long term vision?
“We want a world that encourages trust, adventure and intercultural connections.”

Are we sure about that? I guess if we want people to find rather like-minded ones then we should very quickly spread the news that we are actually not a hospitality network - which in my opinion is the public image of Trustrotos, and that we want a world that encourages that you stay with the people that you already know, where you are safe and don’t have to learn anything new :wink:
I would love to talk more about that, too :slight_smile:

I do nowhere (and will not) conclude that the size of the network correlates with the user experience, this is exactly what I wanted to express in my first 2 sentences. I just think that encouraging people not to register here if they do not want to host anyone will tremendously increase the quality of the network. Or at least install a mechanic that sets your profile to “can not host” if you are not active for more than x. Any way to differentiate between zombies and people that want to host.

Until today I saw all the little unpersonal green and orange dots as some of my kind and thus the search for a host became is hard when you are somewhere with more people (yes, I know that this is already addressed in this forum).
Following the words of @mikael, looking at the map with all those unpersonal dots that are not of my kind disabled, the map of course becomes clearer. making me want to have more and more members (but not somewhere, where already many are present). Also the whole idea of trustroots that I had in my head lost a lot of quality :frowning:

Now in for spreading faster it would help to strongly recommend Trustroots for it’s feature (that has then to be implemented) that you can be part of a specific tribe and not give a fuck about the rest (e.g. only be seen by tribe members and trust that everyone stays well-behaved within their tribe). A little bit sorry for being cynicyl here: Yes, have all this hippies/vegans/cyclist/families go to hell (they are all the same anyway), I only host pure(!) AcroYoga people and hitchhikers.
Like really, this is a feature, not a bug. But then it has to be sold as one.

The bug is, that I can not trust all hitchhikers. And that I can not trust all cyclists. And not all vegans. [¹] So I don’t get why to throw then all into one bucket…? How to differentiate between people that are from my tribe that I can trust and those that are from my tribe that I can not trust. Especially if there are very many (that all very honestly clicked only the tribe that they belong to)… should we then have a trusted-trustrooters network within trustroots?

One thing that I see through the support platform is that there are people doing public work about Trustroots that we could pamper a little more. People that blog about Trustroots. I try to be very aware about that on the support platform and am happy to be talking this people. For now it is just hard to find myself through all the different ways to access information about Trustroots. I didn’t know about grafana.trustroots.org until now. Is that public?

Also, the blog (ideas.trustroots.org) could be more prominent on the main page and could contain a “What happened this month/quarter” section where people see what happens behind the walls - without clicking and reading through all GitHub and Meta Topics. But it is of course quite some work if one is not involved in everything. Could this be automated? X GutHub Topics finished, X Support requests answered, X posts on the Meta-Forum…?

[Âą] (notso)funfact: I recently met an extremely agressive, alt-right, conspiracytheorist, barefoot, vegan cyclist through trustroots, quite a shock but what did they do wrong besides talking bullshit, who am I to judge them?

Thread about tribes [was: Tribe suggestion: language enthusiasts]

After sleeping some nights over this thoughts I wondered:
We could approach already existing communities asking them to host a tribe here on trustroots. That might require them to administer the tribe and head over the responsibility of the tribe-user experience to that group that hosts this tribe and might require a clearer description and rules-arrangement but that way we could access bigger amounts of active users that come as a group. They might most probably be not open to new tribe members and might need some digital safety and privacy but that would be a really good feature to sell the idea of trustroots to them.

For that the structure of tribes would have to be different, I added some thoughts in this post:


I think the various dots should fade with last login, and the time to reply.
If someone hasn’t been active, or gotten any requests, a test e-mail could be sent out, which the member can reply to(?) Maybe with some tips for improving their profile.

Could also additionally do it with ratio of 10 last positive answers, but that isn’t recorded (?)

There is also the option to have users select dates and time-frames they can definitely host, but it gets invasive and problematic quick if done wrong.