Starting a new thread to develop ideas for a new keyword model.
I liked the original planned update to include both a staking and consumptive model but perhaps tweaked more towards consumptive.
Staking is a very interesting and innovative way to tie up PRE and push advertising or links on certain topics of interest. I also think the PRE tokenomics benefit by having these staking utilities that take tokens out of circulation even if only for a period of time. Speaking of time periods it may be beneficial to implement an unstake cooldown period eventually. Perhaps if you choose to unstake keywords there may be a 24 hour or 48hour or 7 day or 30 day cooldown period where you still get the benefits of the stake but once the cooldown has past you would be able to unstake and remove the tokens. I don’t think the a cool down is needed at this time but eventually it might help Presearch understand risks better. Bottom line I want to see staking for keywords continue.
However, the problem with current keywords is in 3 primary areas:
-
staking is currently winner take all. This drastically limits the number of stakes and staked tokens that would otherwise participate because very few would choose to maintain a stake if they are not the highest bidder. The first change should be to implement a winner take most approach with all stakers of the keyword sharing the remainder of views shown in the proportion of their stake.
ex: 4 bidders (10,000; 5,000; 2,000, 500 PRE) The 10,000 stake as the highest stake would get 50% of the total staking views and 57% of the other 50% prorated views for total of just over 75% of total staking views. The 5k stake would get 28% of the 50% prorated share of views. the 2k stake would get 11% of the 50% prorated views, and the 500 stake would get only 3% of the 50% prorated views. In terms of total views on a keyword with 100k views a week the highest staker would get 78,500 of those views and the lowest staker would get 1500 views out of the 100k views. If the winner getting 50% immediate boost is too much that could be adjusted to 30% as a bonus to being the winning staker this would allow all other stakers to share 70% of total staking views getting more for their stakes. evenutally a fair winning stake bonus will be figured out to maximize value to advertizers and total keyword PRE stakes. -
There is no consumptive option. As I said above I think keyword staking is a great USP but does not bring any income to the project although still beneficial to the project and token especially when more users and advertisers are looking for ad space, as this has the potential to lock up many millions of tokens. Consumption on the other hand would put PRE or fiat $ back into the treasury as income. What you could do either simultaneously or after implementing #1 above is add a consumptive keyword option. I think initially 70% of total keyword views should be allocated to consumptive keywords. This would work just like the staking model in that winner takes most and all others consumptive bids would get views in proportion to their consumptive bids. Again what the starting bonus is for a winning stake or consumptive bid is up to the community and team for ease of math I am using 50% bonus to winners.
When you put it all together you have staking assuming no consumptive bids winner would take most and the remainder in proportion to their stake ie: 1 stake and no consumptive bids = 100 percent of views on that keyword; many stakes = highest stake gets majority of views say 50% then the remainder 50% is split between all the stakes in proportion to their stake. This allows the little guy to still compete and get attention. The moment a consumptive bid is placed on a keyword they would immediately get 70% of the total views on the keyword regardless of their consumptive bid amount. This prioritizes consumptive use (paid advertising) as even a nominal bid would outperform hundreds of thousands of PRE staked (this may entice the staker to enter in the consumptive bidding process). While still making staking now only 30% of total views still winner takes most and prorated remainder worth maintaining their stakes. Initially this model would entice small bids for great results and eventually make consumptive bids more competitive with user growth or on highly productive keywords. If there are multiple bids for consumptive keywords the winner takes most and other biders are given views in proportion to their bids. Ie: stakers are getting 30% of total views and consumptive bids are getting 70% with winner take most 50% of the 70% and all bids prorated views of the other 50%. In terms of total views with the same 4x stakers and 100k view assumptions the top staker is now only getting 23550 views, bottom staker is getting 450 views, and the sole consumptive bidder in this case even if only bidding .01 PRE per view would now get 70k of the total 100k views for a total consumptive cost of 700 PRE. When a user chooses to bid a keyword consumptively they set a total advertising budget in PRE for those keyword(s) bids and that PRE upon confirmation will be locked and used in accordance to the views paid for. If the budget in the above example was 1,000 PRE (10 PRE may be the initial non-refundable upfront cost to prevent gaming and overuse) after the 70k views 710 PRE was consumed (10 non-refundable fee and 700 pay/view) paid into the PRE treasury. The advertiser can at anytime cancel their keyword bids and get back all remaining tokens except the non-refundable amount which is paid upfront upon confirmation. If there are multiple consumptive bids it would work the same as the staking but bids now share 70% of views while stakes now only get 30% of total views.
This model would sustain staking and make even nominal stakes or bids valuable and eventually competitive when number of regular users increases. -
Not easy to mass update or easily track keyword stats. I think this could be solved easily with the above changes but now that staking would no longer be be winner take all you could submit a stake for multiple keywords at the same time say 1000 PRE x 10 keywords or whatever amount of keywords you require = 10,000 PRE staked and every change to that ad would update all those keyword stakes immediately. This is a simple solution that could be implemented quickly and easily but admittedly is not the optimal solution. I think a more optimal solution is to be able to set a stake limit order whereby if no one is staking on any of the specific keywords you designate it would only use the minimum 100 stake and would bid up automatically against other advertisers to your max stake amount per keyword for those designated keywords against other competitor advertisers in 100 PRE increments. This would make keywords much easier to update in mass and set it and forget it instead of hundreds of individual tedious updates you can now possibly bin them into low, medium, and high value keywords and provide max stake for each category updating hundreds or thousands of keyword stakes in seconds. If you set a max stake for 100 keywords prior to confirmation it would calculate the total current stake amount of PRE required by checking against highest stake of your 100 keywords and leveraging up to your max stake amount on each if required and 100 if not staked. You would then select the amount of PRE above that current total as your pool for bidding upto your max. Upon confirmation the minimum + the bidding pool would be locked. You can always unstake whenever you want or stake individual keywords. You can also choose not to do a max stake and instead just stake a specific amount for as many keywords as you wish example Presearch would just have an initial stake amount input box and a max stake amoutn input box if you only wan to stake a specific amount like 100 PRE on 50 keywords you would only input an initial stake amount and leave the max stake amount clear. Upon checkout the PRE stake lockup would calculate to 5k PRE. If you wanted to make staking have a cost to cover development of this new model you might charge a non-refundable fee for the stake say 5-10 PRE so that when you confirm the total required is 5010 of which 10 PRE goes directly to treasury and the 5k is staked on keywords as described above. This way there is no hidden cost or pool other than the stake that would need to pay anything. (And you can pay me 1 of every 10 non-refundable PRE if you implement this cost structure for staking and consumptive as payment for coming up with this way better and very usable model ;^)
To better manage stakes an API similar to the node API could keep track of the number of views of all your keywords and apps could be created to track and manage their performance. Within the keyword page a new column should be created for views and clicks. This would be in addition to current total views and clicks. The idea is that you could see both your total cumulative views but also if you change a stake or the ad or wanted to run week over week comparisons to different ads to compare performace you have a sort of trip-odometer-widget for views and clicks that track from the day and time of the reset without resetting or clearing the totals. An API and app would be able to do this best but until then this is an idea to make it more useful and advertiser friendly.
//
Why both staking and consumptive. because this is a unique selling point of advertizing on Presearch that should never fully go away and because the number of active users is so small. This would help tokenomics and probably easily triple or 10x the number PRE locked up in keyword stakes (both because of ease of implementing large amount of stakes and changes but also because their is still views even if not the top stake). This model also gamifies consumption in proportion to the value given to the advertizer. If the advertizers are getting plenty of value they would continue to consume more and get more value and the market would self regulate for competitive value. but if the value is still not here yet due to low number of current active users staking would still be a great option locking up way more tokens possibly getting new non-refundable fees and creating value for all. When there are no consumptive ads the stakers get to share 100% of the views on their keywords. With the number of PRE in circulation increasing this type of lock up utility is great to have. If you start charging PRE to stakers on the other hand as has been suggested by the team and the value is not here yet for most of the stakes users would simply unstake and sell thier tokens or unstake and hold or move PRE to nodes or search staking. Turning a no cost value proposition or even modest income potential from non-refundable fees into a higher cost (if tokens move to search or node staking) or sell presure to the token. I think once this is built out you could play with the percentage maybe it would need to be 90% or even 95% of total views to the consumptive ads but eventually the community and team would find an equilibrium that works.
