what does a good cluster look like

What Does A Good Hotspot Cluster Look Like?

40-60 Hotspots, all about 1,500m away from any other, with any one of ’em having clear line of sight to at least 30 others.

Wait, you want more, and you want to know how I came up with that (with help)? Start by reading my post on PoC cycles, otherwise you’ll be missing some key points. Done? Great!

Let’s start with the worst case for HNT earnings: You’re a lone wolf Hotspot, with no Hotspots within tx/rx distance. You can’t Witness, so you miss out on 75% of available earnings. Even with a few Hotspots around you, the odds that any of your Hotspots will Beacon and initiate the larger part of the earnings cycle is low. Short version: You’re not going to earn very much.

The next worst case is the opposite. You’re surrounded by thousands of other Hotspots who are overcrowding the hex density rules and no one is earning very much because all your earnings are scaled. Again, you’re not going to earn very much.

Ok, so now we’ve established the “my porridge is too hot” and “my porridge is too cold” sides. Where’s the happy medium?

For help with wading through the data, I turned to the #gigs-and-bounties channel on Discord, though Helium.Jobs is another great option for getting help with anything Helium related. In this case, the folks over at LongFi Solutions responded to my request and we started into it.

Let’s start with my hypothesis (which was wrong, by the way). Here’s a decent way of thinking about it, it was just missing enough data points.

I thought the minimum number of hotspots you’d need in order to reliably witness beacons and therefore earn the majority of PoC rewards, would be around 300, just like Albuquerque. Here you can see the 24 average HNT earnings for any given res 8 hex using HeliumVision.

The average monthly HNT earnings of a Hotspot in Albuquerque is around 14 HNT. That may not sound much, but it’s higher than anything else I could find, by a lot. I spent some time centering the map in Hotspotty on cities and seeing how many hotspots were there as well as average earnings.

It’s a fast (well, slow because you’re asking the map for a TON of info) and dirty way to get a “broad strokes” overview, although as you’ll see, it allows you to miss some important points. Here’s what I jotted down.

CityNumber of HotspotsAverage Monthly Earning Per Hotspot
(Greater) San Diego3,8007.41
Los Angeles13,9008.65
San Francisco6,8487.23
New York106645.8

I figured that was a good start, but knew I should probably ask around for help with data that was a little more carefully assessed and parsed. Plus, where was the gold?

Where were the best spots in the country to deploy, and how many Hotspots do you REALLY need? Teling me that San Francisco is not a good place to deploy a Hotspot really isn’t that helpful.

So, here’s what the folks LongFi Solutions found when they dug in to my question.

What is that saying? What the heck is it? On the Y axis is the 30 day average of HNT earnings per Hotpots. On the X axis is the number of Hotspots in the city. We’re zooming in progressively on each graph.

Caveats: I asked for the data on the top 500 US cities. I know, I know, I’m not a statistician. There are probably a ton of things you could get wrong by misinterpreting the data, and among the things not accounted for (off the top of my head) we’re missing topography and signal strength/data size (spreading factor) changes around the world. Still, a few interesting points popped out.

Wait, you want to play with this data yourself? Cool, please share what you find here so we can all learn more! Download the data here, it’s a 32 MB file, just FYI.

Here are my takeaways:

  • You probably need at least 40 hotspots to be earning well.
  • 40 hotspots isn’t a guarantee, it’s just a reasonable target.
  • Once you get beyond 300 hotspots, it’s not that you can’t earn, it’s that avg earnings are low, which is important for fleet deployers.

Here are additional takeaways from the LongFi Solutions crew (specifically, @the-wildcard on Discord)

  • It’s interesting to see how things converge as you move to the right, I guess it’s to be expected though
  • With the pace of new hotspots coming online, a lot of these numbers include hotspots that haven’t earned yet. Will revisit in a week or 2 to see what’s changed.
  • Most interesting to me was that the cities with the highest average earnings per hotspot have 40-60 deployed, a little lower than what I’d expected would be needed to maximize

A few additional points are worth mentioning. Remember, I was looking for best areas, not necessarily the best locations. Locations are straightforward: Get the thing high with clear line of sight and view to lots of Hotspots. Area is harder; how many Hotspots do you really need?

Now we’re all one step closer to finding out. If YOU are a data geek and would like to contribute, please let me know what you’re thinking in the comments, I’d love to include you in the collective wisdom!

If you’d like to join a group of like-minded individuals when it comes to learning about and optimizing Helium deployments, check out the Gristle Crüe!

Until next time, rock on!

10 thoughts on “What Does A Good Hotspot Cluster Look Like?”

  1. Not so much forgot as “didn’t include”. Working with fleet deployments lately, they need to start “big picture” and once they find a likely place, explore further.

  2. Hi Nik,
    Can you clarify: “You probably need at least 40 hotspots to be earning well, and by earning well I mean as of Oct 12th an avg monthly HNT earning of 14 or above.”

    Does that mean you need to be WITNESSING at least 40 hotspots to be earning well? If so, is that measured by the number of “recently witnessed” hotspots in Helium Vision?

    Some of my stats:
    HS 1: 15 recently witnessed. Earning .39 HNT per day
    HS 2: 25 recently witnessed. Earning .53 HNT per day
    HS 3: 28 recently witnessed. Earning .28 HNT per day
    HS 4: 15 recently witnessed. Earning .25 HNT per day

    Those are all in a metro area with around 150 other hotspots. There are no other hotspots in the same hex as mine at resolution 8. All have other hotspots in some of the adjoining hexes. All are at rewards scale 1.

    HS 1 & 4 witness the same number of hotspots but their earnings are significantly different. HS 3 witnesses more that HS 2 but makes significantly fewer HNT. I know it’s more complicated that just witnessing hotspots, but what would be causing these differences if the number of witnessed hotspots is 75% of the rewards and what can I do to improve rewards?

    These are all Bobcats with the stock antenna inside a single pane of glass. I’m planning on changing to outdoor antennas 15-20 feet higher than they all are now. I assume this should increase rewards significantly, or is there more I should do?


  3. Hi Brad, 40-60 is just a starting point. Take a look at the scatter plot graphs to get an idea of where the avg earnings in a cluster of 150-ish is. Getting your antenna outside and higher almost always increases rewards. Rock on!

  4. HEY so thanks for all the great info. good pictures. And answering so many questions all the time. My biggest hang up right now is what cellular modem to use I see the rut used a lot which I read it has option to use VPN on it. Are you guys using them? Because without having a static ip address which not all cell providers have them… so does VPN keep it from being relayed? Then what type of sim are u using what brand and is it unlimited or special iot m2m type or just a cheap 20 or 30 dollar unlimited data talk text with or without static and what brand? Anyone with a handful of set up on cell modems using a family plan let me know please.

  5. Question? … 40-60 in what size of an area? In the article, you did it by cities, but what does that mean area or hex-wise?

  6. Depends on local topography. 40-60 that can *generally* see each other seems to work well, although lately I’ve seen clusters as small as 20 where at least one Hotspot was up in the .2 HNT/day (double the global avg).

  7. Hi Nik!
    I live in a quite hilly area and I found a good spot which isn’t scaled and the nearest hotspot is 30km away with no line of sight. I was thinking about placing 5-6 hotspots in that hilly area among villages which have line of sight of each other. Do you think like 5-10 hotspots will be enough to earn very well if they all see each other?
    Thank you for your help!

  8. Hi Gábor, that’s along the lines of what the Network is looking for (correctly placed proveable coverage) but with 5-10 you may not have as many beacons issued as you’d need to earn maximally. I’d think about how you use the Network as part of the ROI plans.

Leave a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Scroll to Top