Impression Mining

TL;DR Impression mining allows anyone to become a publisher brand / micro-influencer, paid by Index Coop


Index Coop sells product by engaging prospective customers and leading them down the acquisition funnel. In order to do so effectively, we attempt to distill our core values into messaging that resonates with our target customers—and then meet them where they congregate.

Despite the ability to effectively message our value–unless we have cost-effective, high-volume channels to acquire target customers–we will not grow.

Index Coop is limited in traditional large-volume paid channels such as Facebook, Google (Youtube), and Twitter advertising platforms. The GWG is testing niche, crypto-friendly paid channels (Coingecko, Brave Browser, Decrypt, etc) and at the newsletter level, however it Is unlikely these will effectively scale with the Coop’s needs. Because of this, Index Coop must lean heavily on the few channels that have potential for cost effective, high-volume acquisition.

Index Coop has seen its greatest success marketing at the social level, with organic Twitter marketing and with viral content—often leveraging a distribution partner such as Bankless. Announcing and trumpeting Index Coop and product achievements, use-cases, promotions, et al generates several hundred-thousand impressions per week across official and supplementary accounts. To date, these activities have produced astounding ROI as the costs are generally comprised solely of labor.

The motivation behind impression mining is to try and incentivize a widening of the number of accounts participating in Index Coop promotion. The aim is to determine if volume can scale without diluting cost-efficiency too much.

No cost // Small funnel (x00,000 Impressions/wk) → Small cost // Wide funnel (xx,000,000 Impressions/wk)


The program would seek to treat any and every account as if they were a certified Index Coop influencer/publisher, and to ascribe an open ended CPM (cost per 1000 impressions) for certain media types:

Social Impressions (e.g. Twitter) - $8 / 1k impressions

Editorial Content (e.g. Blog posts) - $25 / 1k reads

Video Content (e.g. Youtube, Twitter Video, Tik-Tok) - $50 / 1k views

EXAMPLE: @DCinvestor brings receipts from 11 tweets about IC or its products from March 2021 totaling 240k impressions; he is rewarded with $2.4k worth of INDEX.

Participants will simply snap a screenshot of their impression counts and submit to the portal for a monthly stipend (paid in $INDEX). Judging submissions is at the sole discretion of GWG. The reviewers reserve the right to cap a payment or deem an account ineligible for current or future rewards.

The program will not tolerate/reward:

  • Spamming followers
  • Paid impressions
  • General trickery
  • Inauthentic content
  • Misleading content
  • Saturation of themes


  • Unless you have built up an audience beforehand, consistent success with this program will not generally come swiftly.

  • Any Coop member receiving a pre-determined stipend by the Coop or on a vesting program is not eligible for rewards

  • Guidelines and workshops for effective posting/publishing and brand-building will be held and shared. In general, content should aim to be:

    • Shareable
    • Aesthetically pleasing
    • Distribution-minded
    • Uniquely Insightful (Use-Case, how-to, history, stat/metric, story, testimonial, infographic, video)

Next Steps:

  • Feedback on parameters
  • Kick off program internally
  • Content workshop
  • Post guidelines to Gitbook
  • Impressions Submission portal
  • External participants

Hey @LemonadeAlpha, I think this is an awesome idea.

Quick question, do you think we should be trying to link this to quantitative product outcomes like TVL, unit holders, etc.?

Basically, what I’m wondering is, let’s say there is a lot of engagement and we do pay for it. But if we are not seeing growth in TVL or unit holders, is it still worth paying for? And then also if we do see growth, can we attribute it in any way?

Would love to hear your thoughts.


This is an amazing strategy! I looking forward to seeing the results.

Just some considerations I wanted to highlight:

  1. How will you regulate this?

My suggestion would be to create parameters for each content type and also limit the amount of influencers per content channel.

  1. We should limit the amount of rewards an influencer can earn per month and create sub-budgets for each content channel. According to Proposing a Growth Working Group (Pt. II) we only have $20-$50K to spend on this strategy for about 3 months so about $6.6K-$16.6K per month.

  2. I also share @verto0912 concerns with how we can track correlation with impressions spend.

I believe this will provide a lot of valuable insight on how we should structure the Impressions Submission Portal.

1 Like

I like that this can help reach more people and introduce Index products.

My two thoughts are:

  1. that the next steps go well, they’ll be key.
  2. would the content need to flag that it is impression mining in any way? It seems close to sponsored content which platforms are very careful with marking, and social platforms have rules around. The mechanism may not fall within the definition but public perception of Coop related info could be affected. Done well, I don’t really see any trouble but does anyone have any further info or thoughts?

Great idea with one question re ability to implement. Seems this implies a pretty heavy editorial monitoring and approval process happening after the fact (as I’m not sure we can act as editor before people independently post on their Twitter account)

Would you propose scanning submitters content portfolio at month end before divvying out rewards @LemonadeAlpha ? Seems that’s about the only scalable way to do it - and even then it’s quite manual.

Other than that, I guess we use our BD CRM to reach out to influencers on Twitter who’d be interested in using such a program.

1 Like

I think at this stage we’d need to be focused on impressions until we’ve built out that level of tracking infrastructure @verto0912.

Hopefully we can get that built out soon, to enable this sort of tracking and the referral program too.

1 Like

Hmm. I know that DC likes the Coop for sure. But one of his last post about index was missing his critical thoughts he always has even on very good things. But I hadn’t any deeper thoughts on it because because he was absolutely right about the coop. So. Okay.
But in my opinion something like this should be flagged. Maybe I see it a little bit differently because laws over here in Germany are more strict.
And for a practical critique, maybe this has to go a step further. More direct into dark social. Like Adidas is doing it now for years. Keep in mind, most trusted information are still provided via private communication.

1 Like

Thanks Verto this is a great point and something I’m thinking deeply about.

Programs like your Zerion proposal and the og referral program are awesome because they allow for directly attributable, contingency-based sales to be rewarded. The problem with them comes down to scaling, given participant preference and the trouble of absolute attribution.

My hypothesis is the pool for people who want to participate in a program that only pays if there is an attributable sale is much smaller because direct attribution is not fully inclusive (If I see your ad and find my way to DPI somewhere other than your link, no reward) and because the aesthetics of promoting a referral link are suboptimal relative to the flexibility of posting in a purely promotional manner.

The other problem is, iiuc, we can only attribute purchases that come through one of our known venues ( / TokenSets) and the vast majority of trading activity happens on Uniswap’s UI.

Like advertising before attribution-tech, we’re going to have to blindly trust our funnel to some extent. At a high level, we know that if we find our target audience and relay targetted messaging, there is a way to find DPI/TTI/MVI etc (Uni, Sushi, aggregators) and to buy it (is liquid).

This program seeks to trust that the funnel is sufficiently optimized (but of course can get better) and that if we can feed more input into the top, it should trickle down. Of course, we must monitor this to see if effective. If we see a huge increase in impressions, but that is translating to anemic growth, we’ll need to do further analysis. Effectively, if we open up this program to 8-digit impressions weekly, I’d expect to see tangible uAUM/uSupply growth relative to before the program.

@Mringz and @DevOnDeFi Hear you guys on the regulation front. That is the piece that gives me the most anxiety. Ultimately, my hope is that decisions will be common sense. Would be open to hearing any thoughts on how to harden this.

Devon you are correct in identifying this as manual and after the fact. Ideally this is a day for me and/or someone else to go through and make judgements.

On limits, I think it could make sense to limit in the short run but the goal is that this program can scale pretty widely. One thing I forgot to mention which would be communicated on the public memo would be a minimum amount of impressions for eligibility (e.g. 1k per submitted post & 50k per month).

@mrvls_brkfst I think we could figure out a way to flag it, but the goal would be to try and balance the idea that we do this to support those who might organically or with a light push be willing to promote something they already like and/or use AND that it is paid. The consideration being, we want the participants to feel organic and not feel like they’re paid shills, and we want the audiences to feel that sentiment to. Identifying ways to tease this out is important and ongoing.

1 Like

Great post and I am excited for this program! Couple questions:

  • How will we handle threads? For example, I’ve had success with threads in the past, but impressions vary on each of the individual tweets. Would threads be analyzed based solely on the top tweet’s impressions or would each tweet be eligible?

  • Could you provide some guidance on what you mean by saturation of themes?

1 Like

Thank you!

How will we handle threads?

  • Each tweet in a thread would be eligible (and would count towards one single minimum of 1k impressions)

Could you provide some guidance on what you mean by saturation of themes?

I wish I had a better example but honestly I haven’t seen it too much and just getting ahead of it. This would mean that if certain data points were becoming exhausted (purely subjective), we’d want to taper down on them.

Something I’ve been thinking about which is potentially better than just a warning is that we could apply multipliers for certain content (e.g. for April MVI gets +20% impression multiplier, or for May TTI gets -10% multiplier)


Thanks @LemonadeAlpha agree with you. I think at $20-$50k for this program, the budget is low enough for us to experiment with. Then analyse & iterate :+1:


Absolutely! A tight-rope worth crossing.


Really excited to see this experiment play out! Let me know if there is anything I can do do help

1 Like

This is great. I think we can worry about content issues as they come up, no reason to make perfect the enemy of the good here


Great concept. Excited for this to move forward.

1 Like

One of the rare but absolutely game changing ideas if it works. True innovation here!


Great Idea. I am on board! Time to write some blog posts. :slight_smile:

1 Like

lolol… just found this 1 funny :speak_no_evil:

Hello guys, this is a post for feedback on the Impression Mining of IndexCoop.

I got decent experience in crypto marketing on Twitter, 4chan, TG, Tiktok, mainly with another Top500 Coin I was involved in. Hence, I am questioning the effectiveness of this Impression campaign.

The reason: the program is not suitable for beginners and also not for (big) Influencers.

This post will mainly focus on Twitter marketing. First, I will explain the impression numbers and rates/followers, the average CPI (Cost per Impression) and why beginners and experts are equally priced out. This post will show that the average contributor will not have the reach to earn enough to participate in the long term with ImpressionMining. Also, Medium numbers and “big influencer” participation will be discussed.

In the Orientation Call Slides (New Joiners Call) the Impression Mining was mentioned. Because in the New joiners call total new people to Crypto attended, I decided to give you straight feedback for this program.

The given numbers following are from the Orientation Call Sheet (New Joiners Call):

Social Impressions (e.g Twitter, LinkedIn) => $8 /1000 impressions
Video Content (e.g Youtube, Twitter Video) => $50 /1000 views
Editiorial Content (e.g Blog posts) => $100 / 1000 reads

Example: Tweet post on your wall/feed
Sources consider 20% Impressions over your followers to be good or average. Why? Some Twitter followers just are not logged in, inactive or duplicates.

Let us consider you have 1000 followers. Rarely you will manage to get over 1k Impressions just by tweeting in a standard fashion (with a 20% rate). You will make an average of 200 Impressions with your classic tweets.

Let us assume the Twitter Account participating in Impression Mining has exceptional knowledge about Crypto Twitter and knows how to engage with the right hashtags or in the proper threads. Maybe then, he gets here and then some four digits Impressions.

One Tweet of mine from a low follower alt account got 4k Impressions for tweeting “Sell all”. Nothing else. Just “Sell all”. I engaged in a high emotional thread started by a top account.

So some accounts will reach over 1000 Impressions, I am sure of that. But likely not with quality tweets or even megathreads (a megathread is a multiple tweet cascade focused on high-quality information and summaries).

Even if they reach four digits impressions, they will barely get any money with it. They will get 8$ / 1000 impressions, and maybe reach few times the minimum amount of impressions needed to get paid.

In Medium this situation is barely better. Contributors will get $100 for 1000 reads.

A read is calculated when that viewer scrolls down till the end of your Medium story. Shorter articles have a higher view to read ratio. In other words, writing smaller articles on Medium is incentivized to maximize their read number. Medium is defining: “A good Medium read ratio is generally between 20 and 50%”.

If we calculate that, Medium contributors need to get 2000 people to click on the article in the BEST case. In the worst case, they need 5000 to get $100. Again, I do not think the average crypto account has this reach or force to make this happen - especially not often.

Big influencers:

The same happens for big influencers with >10k (real) followers. Daily Hodl shows here greatly that 30 twitter influencers are responsible for 70% of all $Crypto mentions. Why is that important? I want to know how many of these 30 Top Influencers or similar ones - are participating in your Impression Mining Program.

An account with 1,000,000 followers is also bound to the Impression rate mentioned above. If he has 20% on average for his tweets - he has a Tweet Impression of 200,000. This equals to 200x8$ = 1600$.
In the 2017 UK eMarketer survey, marketers said they would pay £1,351 ($1,779) for a micro-influencer tweet, and as much as £64,798 ($85,339) for one from a celebrity.
The rates for >1M Follower Crypto accounts are likely higher than 1600$. How much, I can only speculate. If I’m wrong, please let me know.

We just saw that crypto beginners and even the average crypto Twitter people can barely reach enough followers and also not enough views to participate the ImpressionMining in the long term.
Also, big third party influencers are (presumably) not shilling for IndexCoop. The market prices are just higher.

If your goal is to do social-media advertising, you can use this money and hire a top agency with the same money. The cost of 1000 market impressions is around $6.5$ (Webfx 2020). That means from a market perspective IndexCoop pays more. Even if I think Crypto impressions are more worth.

You could also use this money to hire selected top influencers. Because in reality, as in my example above, I don’t think many third party crypto Twitter accounts will participate in your Impression Mining.

You can also change the system fundamentally. To have a solution for your whole program I don’t know enough details and I’m just a guy from outside.

At this stage, I would not attend or recommend ImpressionMining. There are just better alternatives around to earn money in Crypto for beginners and even intermediate people. Yes, with dedication and good work you can do anything in life. But maybe not this one. But maybe you can prove me wrong here. I like to be wrong, so I can learn something. :blush:

Best, hopefully soon Bronze Owl :owl: