Proposal: New App Reviewer Buffer

[polluterofminds]

What is the problem you are seeing? Please describe. Social media marketing is a key function of successful companies, and we should be embracing that in the decentralized web community. However, there are significant problems with incentivizing "reach". The current App Reviewer, Awario, as best I can tell, does not have a component that tracks engagement or quality of reach. In tracking just reach, app developers and the Blockstack community gain very little value. Reach is essentially meaningless when it can be gamed and when it does not generate leads.

In fact, Awario is more of a tool for monitoring mentions not increasing engagement. This is misaligned with the goals of App Mining and Blockstack.

What is the explicit recommendation you’re looking to propose? I propose a new App Reviewer to handle social marketing. I believe that App Reviewer should be Buffer. There are a few things that will really benefit the entire process here:

-Buffer has a tool for scheduling posts. So one requirement could be that all social marketing scores be based of posts that originated out of the app's Buffer queue. -Buffer has an analyze tool. This tool would need to be evaluated, but it appears to track the quality of reach and engagement. -Buffer encourages apps to actually put in the leg work rather than resorting to slimy bot farms.

Describe your long term considerations in proposing this change. Please include the ways you can predict this recommendation could go wrong and possible ways mitigate.

Buffer may not be interested in participating in App Mining. Buffer may not be willing to easily share numbers from a user's queue. Devs may be adversely affected because they don't use social media at all now.

[ntheile]

@jehunter5811 ,

Nice, I will have to check out Buffer. I have not heard of them.

Have you heard of SproutSocial? https://sproutsocial.com/ They are based out of Chicago and I know a developer who works for the company.

Sprout Social’s social media management software and solutions help you find, form and deepen real connections with the people who love your brand.

I have not personally tried it out yet but it looks interesting.

[polluterofminds]

I used SproutSocial way back when. Buffer, I think, provides better tools but admittedly I haven’t used SproutSocial in a while.

I specifically propose Buffer because their goals are very much aligned with Blockstack. They are not trying to get companies to promote bots and gain reach. They are trying to help apps gain organic growth by doing the actual leg work of social media marketing.

[polluterofminds]

In case people don’t know about Buffer, here’s the link https://buffer.com

[cuevasm]

First, big fan of Buffer and I could warm up to this.

That said, here are some things I am thinking about right away: I don't think it's very fair (or accurate) tolook at someone's marketing efforts or reach/growth by only looking at posts they send. Other forms of marketing may be WAY more effective for certain Apps and they would suffer for doing the right thing and focusing on those. We're essentially imposing social media marketing on everyone as a requirement to be successful in App Mining, not sure that's the right call. Also, Buffer doesn't differentiate between posts you send from their interface and ones your account sends from elsewhere, it tracks metrics on all the sent posts in your accounts - not sure how we could separate.

Other things:

• Buffers engagement score is opaque
• A buffer agency setup might work to give visibility to everyone on each account's metrics, but we may be looking at custom work here they may or may not be able to do
• This doesn't encourage teams to do the right marketing for them, it encourages them to do social - what I liked about Awario was that it was quite a bit more holistic - a team may be best-suited spending their time and money on PR and nothing else, but they would be punished.
• New apps with no following would suffer.
• There's nothing about Buffer that inherently encourages people to do the legwork, in fact they have an automated tweet sourcing mechanism built in. They can still do all the same bad stuff, I think only counting Buffer posts is the attempted solve there, but see my points above.

SproutSocial is another awesome tool and I use it regularly. I see similar limitations.

This decision kinda comes down to if we want to say to the community, hey, you have to do direct social media marketing to be successful in this program. I think that's right for probably 80%+ of folks, but maybe not everyone.

Related idea but still limited --> https://sparktoro.com/tools/sparkscore Combining a few scores like this from different places into one score might be compelling

[polluterofminds]

We're essentially imposing social media marketing on everyone as a requirement to be successful in App Mining, not sure that's the right call.

That call was already made with the introduction of Awario. If you do not do social media marketing, you will not get any sort of decent Awario score. Not sure what the difference is here.

If we have a social media score (and I know Awario tracks blog posts and news articles, but those are clearly worth less based on what I've seen so far), then using some method that does not encourage paying bots and driving up artificial reach numbers seems to make the most sense. Maybe Buffer isn't the right tool, but I think it's a heck of a lot better than looking at reach.

Again, I think if we are going to use social media as a component in App Mining, engagement should be the criteria, not reach.

[kkomaz]

We're essentially imposing social media marketing on everyone as a requirement to be successful in App Mining, not sure that's the right call.

App mining is most definitely imposing social media marketing as a requirement to be successful. Resorting to twitter bots or paying for PH upvotes was not even a discussion until it started to have a significant impact on the score. As evidenced by PH upvote impact

dmail had 1000 PH upvotes when they initially hunted but somehow only had 200+ users on it.

I'm not saying dmail bought votes but PH having a 63% impact on the final score is an indication that social media marketing is a HUGE factor to be successful in App Mining.

[cuevasm]

Not sure I agree totally, Awario picks up Mentions and therefore Reach from news and blogs across the internet which is MUCH wider than what Buffer would get. To your point, I don't think anyone has been that successful at this so far, but it's actually a huge way to get great numbers. It also picks up Reddit, which is somewhere between social and web.

[cuevasm]

Point being I feel like we shouldn't go more specific, we should be looking to assess more broadly if anything. Maybe it's not possible, but aspirationally, the more complete marketing picture is better than a super specific (1-3 channel) one.

[cuevasm]

I'm generally in agreement on Engagement - though I think the value of it varies dramatically app to app and there are probably many where just raw exposure is ultimately more helpful to user acquisition than a handful of tweets being highly engaged with. A 90% engagement score on a tweet that only 100 people saw probably results in less new users than a tweet that was seen by a million or 2 people with only a 5% engagement.

[cuevasm]

Would Miner's generally be open to an engagement score where the formula isn't published?

[polluterofminds]

I doubt people would be in favor of anything less than transparent. We saw this with the Product Hunt team vote early on in App Mining.

If people believe that it's possible to have reach measured without fake reach being calculated and without an labor-intensive process of vetting actual social reach, then I agree with everything you're saying, Mitchell. I just don't believe that it's possible to have both of these things.

I think we can have reach continue being the measurement and accept that people are going to pay bot farms. Or we can eliminate reach as a measurable part of the Awario score (or eliminate Awario entirely). I hope someone can convince me otherwise, but I don't see it being possible.

[friedger]

I think we should work on marketing guidelines that are valid for Buffer and Awario. Before we don't have I don't think it make sense to explore a new reviewer.

Whether a marketing activity (bot farm, influencers, PH upvotes, ..) is paid or not, does not matter, it matters whether it is within the guidelines.

For me reach is meaningful, if these are real impressions, viewed by people. Ad networks have a good understanding of whether an ad is viewed or not. Maybe we should measure how many people clicked on an ad of our apps? E.g. 50€/month for each app is spent on twitter ads. The reach is then calculated from the ads analytics.

[muneeb-ali]

There seem to be two discussions here:

a) Are marketing effort and social media engagement useful metrics (regardless of how we measure it)?

b) How can we counter attempts to game such metrics?

RE the first, I'm of the view that yes they are. If no one knows about the great app that you've built then that is not really helping towards the mission of Web 3.

RE the second, I think there will certainly be noise and certainly be some unintended consequences of measuring the score (like paid bots). The challenge is how to get better data out. Hacker News does a great job of making their system less gameable. My hope would be that the App Mining incentives would drive people to improve how gameable these systems are. That can be a big win.

[cuevasm]

Hey all, we're going to prioritize testing out the altered Awario models for now and in the meantime see if we can get a good contact at Buffer. This would be a bit hard to do with their off the shelf tool so I'm not hopeful about being able to work with them easily, but we'll see what we can drum up.

[Walterion01]

@cuevasm What will happen in case of founding out some participants did game the system (like Awario incident or paid bots for PH upvotes) in the previous round? It seems logical to effect "Score Last Round" to prevent the effect on future rounds.

[dmailonline]

We're essentially imposing social media marketing on everyone as a requirement to be successful in App Mining, not sure that's the right call.

App mining is most definitely imposing social media marketing as a requirement to be successful. Resorting to twitter bots or paying for PH upvotes was not even a discussion until it started to have a significant impact on the score. As evidenced by PH upvote impact

dmail had 1000 PH upvotes when they initially hunted but somehow only had 200+ users on it.

I'm not saying Dmail bought votes but PH having a 63% impact on the final score is an indication that social media marketing is a HUGE factor to be successful in App Mining.

First time to note this, let me clarify some facts we think that you missed. 1 - 200+ is the number of new Blockstack users because of Dmail, it is not the total Dmail visitors. 2 - At the time of this twitter post, you can find that total logged in Blockstack IDs almost 450 users. 3 - Who said that in order to upvote and like app idea, you should try it? Personally, sometimes I do upvotes for some apps on PH just because I liked the idea, so, the total number of App users is not the right indication for PH upvotes. 4 - we recommend to check https://theblockstats.com/ to find where is Dmail according to the total number of installs. 5 - The biggest part from June score for Dmails came from PH because at the first month of participating in App mining program, you only evaluated from 3 of 4 viewers, this is why you can find this effect. 6 - PH has it's own special (not public) formula to count only credible UV. Last but not least, we believe that the top three apps in ranking may stay under focusing more than others, especially when the App being participated for the first month and get the first place. (like Dmail), at that time people try to think that they did some cheats or some gaming to reach there, but no one looking at the efforts already done.

Please try to review Dmail posts on the slack channels (especially Apps channel) and count the total number of features and update we did announcements about, and do compare with some other (silent) Apps since we have launched Dmail.

Thanks.

[kkomaz]

@dmailonline

I'm pretty active in the community so I'm aware of how many ProductHunt upvotes ya'll had when dmail initially launched relative to the numbers on the https://theblockstats.com and it was around 200. Either way, that wasn't the point of my original response to

We're essentially imposing social media marketing on everyone as a requirement to be successful in App Mining, not sure that's the right call.

Ya'll are doing a great job, and I see all the updates the team is making so I'm not critiquing the development work that's being done. Like I said...

I'm not saying Dmail bought votes but PH having a 63% impact on the final score is an indication that social media marketing is a HUGE factor to be successful in App Mining

The purpose of the reference was that 63% is a huge impact on an app's final score. Having it so heavily weighted in the beginning of an app mining submission can cause OTHERS to possibly game the system via buying upvotes if they are aware of the importance of the first submission.

[dmailonline]

@kkomaz

Every single reviewer in the program ( Except digital rights) is subjective. you will not be able a find the best solution. TrymyUI .... is very subjective, if someone knows about Blockchain or Blockstack, so you can estimate a high score from him if someone just knows how to open his laptop and do some browsing (the majority of the testers in TrymyUI are from this category) so you may get a very bad score. In last round, Dmail got a score of 95.31% from a tester and got 7.8% from another tester .... see the big fluctuation?

Awario ... still we don't have a solid solution to stopping the bot actions.

PH ... if you are lucky at the PH launching it will be perfect, if not then you stuck for a 6 month till you can make a re-launch.

There is nothing very objective and I really like rather than Digital rights which measures what you are doing with Blockstack tools ... very simple ... if you do this, you will get that, 0% to game it.

From my perspective, finding a way to review and evaluate with 0% to game is better than trying to reduce the gaming effect in a reviewer.

Thanks.

[kkomaz]

Yep! Agree with you completely! @dmailonline 👍🏻

Intention of the post was not ill will, (although I can see how you saw it that way) but showing some stats that clearly reveal the short & long term impact subjective scoring can have on app. 🤝

[Arrakis8]

What is the problem you are seeing? Please describe. Social media marketing is a key function of successful companies, and we should be embracing that in the decentralized web community. However, there are significant problems with incentivizing "reach". The current App Reviewer, Awario, as best I can tell, does not have a component that tracks engagement or quality of reach. In tracking just reach, app developers and the Blockstack community gain very little value. Reach is essentially meaningless when it can be gamed and when it does not generate leads.

In fact, Awario is more of a tool for monitoring mentions not increasing engagement. This is misaligned with the goals of App Mining and Blockstack.

What is the explicit recommendation you’re looking to propose? I propose a new App Reviewer to handle social marketing. I believe that App Reviewer should be Buffer. There are a few things that will really benefit the entire process here:

-Buffer has a tool for scheduling posts. So one requirement could be that all social marketing scores be based of posts that originated out of the app's Buffer queue. -Buffer has an analyze tool. This tool would need to be evaluated, but it appears to track the quality of reach and engagement. -Buffer encourages apps to actually put in the leg work rather than resorting to slimy bot farms.

Describe your long term considerations in proposing this change. Please include the ways you can predict this recommendation could go wrong and possible ways mitigate.

Buffer may not be interested in participating in App Mining. Buffer may not be willing to easily share numbers from a user's queue. Devs may be adversely affected because they don't use social media at all now.

My Equipool miner is not working correctly. How do I get it to work? It mines and the miner shows speed as 21 sols/s, BUT when I go to the Equipool data site it shows Speed 0, and Avg speed 0. Why is this app not working? Red letters on my miner say "INFO: GPU0 Rejected share 50ms [A: 0, R:1 (or R:2 or R:3 etc)]. I got screenshots.