Has anyone charted our growth since we moved here? It might be interesting
to do that and then to look back at what was happening at DMS if there are jumps
I know this year our growth has been increasing, but we grew a lot after we moved here,
David @Photomancer posts that information routinely but the detailed data only go back to 11/7/2016. This post here will show you the 2017 data as of five days ago. Here are (most of) the 2016 data.
1 Like
It would be interesting to go back to the move here and to see
the growth since then, then look to see what was happening that
generated that growth, Without some ides, we dont know what to
plan for in the futrue
1 Like
Well, without a detailed chart, I can tell you that we were about 300-ish members when we moved from Ladybird in July of 2014.
There should have been a spike in the months after that due to
te better location
My understanding is that David @Photomancer really only started actively tracking the numbers at the begining of the year. In a few more year’s time, we will have a LOT better concept of what seasonal changes we need to watch out for as well as what the general growth and churn patterns look like.
MakerManager/WHMCS should have a timestamp’d record of every members join date I think
2 Likes
I’m sure you’re right, but it’s not obvious that it’s worth investing all the effort to manually extract and chart that.
If anyone with SQL access is interested it could be as easy as:
SELECT DATE(created_at),count(*) FROM makermanager.users GROUP BY DATE(created_at)
Based on the MakerManager code
5 Likes
How this started. Last year Stan and I were talking about how the previous year there was a large drop in membership from the end of November until the middle of January when things skyrocketed through the end of February. I decided to track this since I was on Finance committee and this obviously is a good piece of information to know.
So on 11-8-2016 I started pulling data from: https://accounts.dallasmakerspace.org/member_count.php this updates in real time for new sign-ups. Cancellations drop off during the night as payments don’t go through - although there can be a two or three day lag/grace period.
I save the data in a spreadsheet, this is what it looks like:
I update this throughout the day. It is always lowest in the morning around 7AM and then slowly goes up, peaking just before midnight. The can be as much as a 15+ drop between Midnight and 7AM, and some days are a plus 15+ (swings this wide are rare). The numbers recorded are for end of the day - meaning my day at the computer.
Occasionally I’ll miss a day … those rare days I’m not at my computer. When this occurs, I list the day as extrapolated based on previous day and following day, I note it in the spread as such:
Note this period covers one of the more volatile weeks.
When I do the calculation for projected income based on that days membership mix there are some adjustments made to Regular memberships: You see the total Regular @ $50 is 17 less, there are 10 grandfather in @$40 (it’s actually a mix of $30 & $40) and 10 that are Life members that generate no current income, I don’t add LT members to the average dues collected, if I did, amortizing on a 5 year period it amount to about +$0.30 to average. But each class is multiplied by the respective dues amount.
On November 8 as the baseline date I use for membership by class changes and income changes.
So using all three of these I get the charts that only nerds can really love:
Membership:
Dues Revenue
I find it interesting to see the changes and try to figure out what the heck is happening. I’ve noted when the trend line goes below the 30 day moving average so it’s easier to see the value. I note the highest value achieved and sometimes the most current value.
So that’s where the data comes from and how it is calculated. I hope it gives some insight into where we are and possibly headed, since just a raw member count doesn’t give a trend on member mix.
If someone will do that and have it create the charts and numbers that would be great!
If it would just pull the number and email at 23:59:30 that would be great also.
2 Likes
So there’s a whenCreated field in the AD database which is part of what we want. I was just looking at this the other day but I hadn’t graphed it yet. Assuming it contains accurate data, this would give us new member signup counts going back to late 2015.
whenCreated.csv.zip (16.4 KB)
I spot checked a bit and it looks like this field might have usable data. Someone graph it and let’s see what we get. This will show us new member counts for each day but I don’t immediately see any other database fields that would show us cancelations which is the other half of knowing our historical daily active member count.
1 Like
Those are the join dates of every member in the AD database. That’s the only field I included for the sake of privacy. It would be great if we also had cancellation dates, and I’d have included that field if we had it, but I don’t see any such thing. As far as I can tell, this is the extent of the useful data AD has to offer us for graphing membership numbers over time.
I did a quick pivot table to count the number of additions per date (without bothering to convert it from UTC to local). The cumulative number added was 3142 (which should also be evident from the number of lines in the data).
If you assume we had 600 members in on July 27, 2015, with those cumulative additions (totalling 4000 vs. the ~1600 current members) it would suggest that for every two members who join, three leave. (and the exact starting number doesn’t really matter - it was just a sanity check on the data).
Those numbers just don’t make sense to me.
1 Like
Given the nature of this dataset, we must assume instead that we had zero. Starting on July 27, 2015 the existing membership began to be added to this database over time. So for some period of time (until all the existing members had been added) the apparent daily join count will be higher than the actual value, but it seems reasonable that this would only last for a few months, making the data pretty accurate by late 2015.
Agree the number leaving exceeding the coming in would mean we are shrinking. It’s actually closer to 3 in and 2 out.
The data strings I’m using from the membership count have no identifying information:
Even if we could just get the very last field “total” associate with a date would be really helpful.
If items above equal people joining each each, if you subtract that from that days total, you get a number that typically will be less than the prior days total, the difference is the number of cancellations (doing this off the top of my head, but something like that). No identifable info.
1 Like
OK, if I make that assumption (which I don’t fully understand but I’m willing to start from that assumption), then there are 3412 in, and a total of ~1628 current members. That would mean 1784 departures and 1628 joins since 7/27/2015 (counting all current members as new joins). That still means that for every one that joined and stayed, another one-plus left. That still sounds like a shrinking membership paradigm.
I must be missing something here because those numbers still don’t make sense to me.
What may not be counted are members that join, cancel/lapse and rejoin. They have the same profile saved in WHCMS - i.e. ID number. Some may be for months some just days, i.e. CC expired. Don’t know if this is the case, but a possibility - but I doubt that would account for that many. So there is something we don’t understand about the data set since clearly we are growing.
1 Like
It’s 3412 joins and 1784 departures. For every two that ever joined, one stayed and is still here today. That’s growth.
5 Likes