Map Improvement Proposal 2 (MIP-2)

Hivemapper Network
Hivemapper Foundation
7 min readMar 10, 2023

--

Hivemapper decided to use crypto rewards because it creates incredibly powerful incentives to align the economic interests of contributors and the network. With the right incentives, everyone is on the same team building a product that will reach scale and generate utility and value.

Over the first several months of the Hivemapper network, some problems emerged with how region boundaries and region weights influence the incentive structure at this early stage of the project. As a result, HONEY tokens are being minted far too slowly. Although the pace is accelerating, at the current rate of 1 million tokens minted per week, it would take more than 75 years for map contributors’ allocation of 4 billion tokens to be issued. Let that sink in: 75 years!

Today, the Hivemapper Foundation is releasing a plan for addressing these problems. The planned changes would take effect in late March 2023, resulting in a larger weekly rewards pool, a more evenly distributed rewards pool across regions, and targeted bounties for high priority regions that will demonstrate the value of the Hivemapper Network to customers.

The Map Improvement Proposal, known as MIP-2, is now open for feedback.

Active members of the Hivemapper community are encouraged to participate by sharing questions, suggestions, and comments in the #mip-2 channel of Hivemapper’s Discord Server. The comment period will be 7 days, running from Friday, Mar 10, 2023 until Friday, Mar 17, 2023.

Three Problems, Three Solutions

Three main problems have emerged with rewards calculations since the launch of the Hivemapper network in November 2022:

  1. Because of the subtleties of how map progress is calculated, inactive regions have caused HONEY tokens to be minted more slowly than intended
  2. The formula for region weight and the size of region boundaries have led to too much variance in rewards between equivalent regions
  3. The highest priority regions have not been allocated large enough rewards to motivate rapid adoption to deliver 100% coverage and consistent refresh.

The Hivemapper Foundation will take three steps to address these problems.

  1. Temporarily deactivating regions with no map contributions since network launch

Hivemapper’s weekly rewards pool is calculated based on global map progress, which is a weighted average of progress in all defined regions. (See here for more details.)

Because it is a weighted average, unproductive regions are a drain on the rewards pool. What does this mean? If a region has 0% progress, it is actually reducing the weekly rewards for all of the contributors in productive regions.

The region weight formula was optimized to support the long-term steady state of the network, when we will see consistent activity in all regions of the world. However, as the chart below shows, almost half of defined regions still have 0% region progress in a given week. Although this percentage continues to fall, this is causing a significant share of potential HONEY rewards for productive contributors to go unminted.

To solve this problem, the Hivemapper Foundation will deactivate approximately 200 currently active regions with no map submissions since the launch of the network on Nov. 3, 2022. This would result in a roughly 25% increase in the weekly rewards pool that is shared among active contributors in other regions, all else equal.

Regions that would be deactivated as of March 8 are listed here. Examples of large population centers that would be temporarily deactivated include Cairo, Hanoi, Jakarta, Johannesburg and New Delhi.

It is important to stress that this measure is temporary. In some cases, there may be contributors waiting for delivery of a dashcam so they can map in one of these regions. The deactivated regions will be reactivated as meaningful contributor activity is observed.

As a demonstration of how region activation works in practice, in March the Hivemapper Foundation will activate new regions that have shown consistent contributor activity, including all of South Korea, the Netherlands and Portugal, as well as:

  • Adelaide, Australia
  • Bahrain (entire country)
  • Belgrade, Serbia
  • Bratislava, Slovakia
  • Melbourne, Australia
  • Sofia, Bulgaria
  • Vienna, Austria

To avoid draining the rewards pool in the future, region creation will now be based on observation of consistent contributor activity.

In this MIP, we are proposing that four consecutive weekly reward cycles with map coverage in an area will trigger the activation of a region, and four consecutive weekly reward cycles without map coverage in an active region will trigger deactivation.

New regions will be based on widely accepted administrative or statistical boundaries. By policy, regions will not be created or edited upon request, but community members may post in the #map-data channel on Discord or use this form to make recommendations.

2. Revising the region weight formula to reduce variance across regions

The amount of HONEY rewards that contributors receive is currently too dependent on exactly how the boundaries of their region were drawn.

To understand how this happens, let’s take a trip to Alberta, home to Canada’s fourth and fifth most populous cities: Calgary and Edmonton.

Both cities have about 1 million people, with a comparable amount of economic activity. In theory, the value of map data from these two cities should be similar, and map contributors from these two cities should get comparable rewards, right?

Unfortunately, it doesn’t work that way right now.

As you can see, Edmonton was drawn with expansive boundaries that include a large amount of rural land. Calgary was drawn with tight boundaries around the city center. Mainly because of the difference in road density, Edmonton has a weight of 0.19% and Calgary has a weight of 0.44%.

It is also easier to achieve higher region progress in a compact region, so Calgary has consistently achieved a higher region progress score. Bottom line: contributors in Calgary have been receiving more rewards than contributors in Edmonton for reasons that are unrelated to the value of their map data.

To solve this problem, the Hivemapper Foundation plans to reduce or eliminate the role of road density in the weight formula. The final formula should also include a base component that is equal across all regions and a variable component based on road volume.

Feedback is welcome on a formula that will balance these priorities:

  • Weight should scale with network utility (i.e. demand for map data)
  • Weight should scale with road volume in a region
  • The inclusion of low-density rural land in a region should not meaningfully drag down the weight of an urban center with a dense concentration of roads.
  • There should be less variance in weight between substantially similar regions

It is also important to remember that the resulting weights are not set in stone. Weights will continue to change over time, by definition, as new regions are added.

The Hivemapper Foundation will also continue to iterate on the methodology for setting region boundaries, including for existing regions, to ensure more consistency in rewards.

3. Launching bounties in five regions and two countries.

The long-term health and growth of the Hivemapper Network requires customers finding its map data useful. If this does not happen, nothing else matters.

To fully realize the value of our collective mapping efforts, regions must reach 100% coverage and then continually refresh the data on a fairly predictable cadence. Otherwise, the region data will only be useful to customers with niche use cases, and those needing to augment and refresh existing data.

This is why it is mission-critical for Hivemapper to start getting regions to 100% coverage. Customers want to see proof that the Hivemapper community can deliver a fresher map than the competition across multiple parts of the world.

To accelerate this process without giving target regions an unfair share of the rewards pool, the Hivemapper Foundation will launch a previously announced bounty service that will allow any holder of HONEY tokens to incentivize map data collection in a specific area.

For the first set of bounties, the Hivemapper Foundation will use its own HONEY holdings to launch five Region Bounties and two Country Bounties.

The first five regions receiving a bounty will be London, Madrid, Paris, Seoul and Tokyo.

In each of these cities, the Hivemapper Foundation will offer a 2x match on normal map contribution rewards, meaning that every map contributor will receive 3x the Map Coverage rewards they would otherwise receive. The bounty will be active until December 31, 2023, and will pay out a maximum of 10 million HONEY per region.

The first two countries receiving a bounty will be Portugal and the Netherlands.

In preparation for the bounty, the Hivemapper Foundation will replace the current regions with a new set of regions for the entire country. For these new regions, the Hivemapper Foundation will offer a 1x match on normal map contribution rewards, meaning that the contributor will receive 2x the Map Coverage rewards they would otherwise receive.

The bounty will be active until December 31, 2023, and will pay out a maximum of 10M HONEY per country.

The 1x or 2x match will only apply to base rewards received from the Map Coverage rewards pool. It will not apply to QA rewards, Hivemapper Foundation rewards, other bounties, or supplemental mapping rewards such as map consumption royalties.

The Hivemapper Foundation is open to feedback on the exact design of the bounties for these regions and countries. Additional bounties for other regions and countries may be announced by the Hivemapper Foundation at a later date.

Once the bounty service is fully operational, it will be made public so that any holder of HONEY tokens can offer them in return for delivering map coverage in a specific area.

Community Feedback

The Hivemapper Foundation welcomes community feedback on the MIP-2 proposal.

To provide feedback, please share any questions, suggestions and comments in the #mip-2 channel of Hivemapper’s Discord Server. The comment period will be 7 days, running from Friday, Mar 10, 2023 until Friday, Mar 17, 2023.

--

--