From 1a94f6e37bb1ebf19d7fa5f53dd3f89aeec15859 Mon Sep 17 00:00:00 2001 From: John Cardinal Date: Sat, 13 Aug 2022 03:30:39 +0000 Subject: [PATCH] removed references to digitalocean from docs in prep for SAAS (no need telling them how to do it without us) --- devdocs/pricing.md | 440 ++++++++++-------- devdocs/todo.txt | 27 +- .../ayanova/docs/ops-install-linux-server.md | 6 +- docs/8.0/ayanova/docs/ops-install-linux.md | 2 +- docs/8.0/ayanova/docs/ops-intro.md | 2 +- 5 files changed, 269 insertions(+), 208 deletions(-) diff --git a/devdocs/pricing.md b/devdocs/pricing.md index 69d75be0..d4babf9e 100644 --- a/devdocs/pricing.md +++ b/devdocs/pricing.md @@ -49,6 +49,12 @@ Subscription hosted (We host they pay a flat fee per month / year to get access Eventually I'm going to want to cash out and sell the business and recurring revenue model is far easier to sell than a perpetual model. +### SINGLE LICENSE WILL CHANGE THINGS + +The tiers are artificial for us because they stratified the sales, when we go to the new system of you pay for every tech / user that you need it will become unstratified and we'll get exactly the size each customer needs. +some higher level ones may want to decrease because there's no 7 user for example, some may increase because they've been holding off not wanting to level jump. +Although in the end I bet it affects new sales more than existing because it's fucking cheap for value right now. + ### Positioning Our price position is in the middle: the best value for dollar, not the lowest cost and not the most expensive fanciest one but best value. @@ -58,6 +64,239 @@ Our price position is in the middle: the best value for dollar, not the lowest c Value to price ration of 10:1 is ideal, means customers feel like they are getting 10 times more than what they pay I've never heard anyone say AyaNova is too expensive +### License types and programs we will offer + +TWO types makes the most sense after considering options: + +#### Perpetual + +- Most similar to current v7 but not exactly the same, subscription is broken out separately +- One time fee, user can use indefinitely +- self installed, hosted and maintained by customer +- least profitable for us long term if they don't buy a maint. subscription +- Without maintenance subscription, eligable for Minor updates only to fix bugs no new features so in other words they buy 8.0.2 they can upgrade to any 8.0.X version release, but not 8.1 as it will be new features added that don't break backward compatibility +- one-time payment, along with the option of a yearly maintenance fee. +- This is basically our current model but we allow upgrades for subscribers +- **HAS CODE IMPLICATIONS** upgrades need to check if allowed based on version number if no maintenance subscription _not_ on date of build. +- TBD - MAJOR RELEASES: not sure yet; if they have an active maintenance subscription can they can upgrade to any newer major releases?? +- If they lapse the subscription they pay a much higher premium but not quite as high as a completely new purchase to reward past loyalty (i.e. maybe 75% or some other % as appropriate of a new purchase or something?) + +#### Subscription license + +- SAAS, pay month to month to keep using it +- We host, maintain and always upgrade they don't need to do a thing but use it +- The subscription model involves recurring payments, typically monthly or yearly. The subscription model can be thought of as “renting” the product instead of “owning” it under the perpetual model. + +#### SWITCHING + +We need a formal declaration of how to switch from one to the other models. + +##### Perpetual to subscription + + they start with perpetual but go fuck it we don't want the hassle anymore so....? + I'm guessing it's just switching to month to month plans and pricing, nothing special except maybe we offer to move their data for them + +##### Subscription to perpetual + + They must purchase a perpetual license as the SAAS license wont' work with perpetual but they can restore their database from the subscription. + +### Pricing for v8 + +#### PERPETUAL + +### Actual price (working out stuff below for posterity) + +**US$135.00 / tech up front perpetual license** +No volume discounts - same price for any size company +Gets you 30 days support and updates and that's it + +**US$100/tech/year maintenance** +Only $8.33 per month (must mention in marketing) +No volume discounts - same price any size company +Gets you one year of support and updates to any release +If it lapses for more than 90 days then you pay full price again to get back on the train (at our discretion so no constant cheap asses doing over and over) + +**UPDATE INFLATION IN TRUE TERMS CLOSE TO 25% so using that** + +IDEA: figure out average license cost over all sizes of existing license sales, factor _that_ with the 25% true inflation and make that the average license price?? + +**update: This concept ok, numbers not exactly what we're going for though** +For active subscribers the average price paid for licenses was 121.25 so with inflation of 19.4 that would be 144.78. +Tech counts by level is single=30 $4,770.00 total sales, 5=85 $11,815.00 total sales, 10=110 $13,090.00 total sales, 15=15 $1650.00 total sales, 20=100 $9,900.00 total sales +(more 10 level techs than any other level and single is kind of lame) +So, if we flat price no volume discount and want the inflation increase, if we stuck with 189 for everyone using the single price that's actually a huge amount for the up to 20 crowd + +The real revenue is in the subscription so a lower-ish license cost is where it's at +But what if no one buys a subscription?? Maybe though, lower up front and no mandatory subscription which must drive some people away, means more sales when you have a sellable product like early days of AyaNova + +This is all about the initial license purchase, not the subscription which I haven't got to yet. Anyway, we were talking about a flat price per license no matter how many you buy. I still like that idea and was doing some calculations to figure it out. If we went that route then the pricing would make more sense to be calculated on the _average_ price per license sold, not on the single user price which would end up being 90% higher for a 20 user site. I used our existing active subscribers as a basis to calculate the average price of each license sold. I totaled the amount we charged for all active licenses then divided by number of techs overall that the represent and came to $121.25 is the average price we charge per license. So with inflation using your calculation of 19.41% that's 144.78 flat rate price per tech. + +Except the current v7 license comes with 1 year of support and updates and I'm thinking of 30 days initial support and updates so basically we're taking away 11 months worth of support and updates which is not nothing either, not sure if that needs to be factored in or not? +The "value" of a support and updates subscription in v7 terms is 35% of initial license price so using the average that's $42.43 a year per tech or $3.40 / month per tech. +So of that 121.25 we are saying that the tech / license portion is 78.82 and the support and updates portion is 42.43 + +**UPDATE CHANGE OF CALCS HERE, PRIOR WAS BASED ON SALES HERE ON OUT WE USE ACTUAL LISTED PRICES** + +Revenue for current v7 users based on subscription only, not new sales and only AyaNova licenses: +Single, 30 active @$55.65 per=$1669 +Up to 5, 17 active @$243=$4131 +Up to 10, 11 active @$416=$4576 +Up to 15, 1 active @$577=$577 +Up to 20, 5 active @$693=$3465 +QBI, 10 active @34.65=$346.00 +So currently $14,418 total revenue for tech subs alone, that needs to like triple just to be worthwhile + +Actual we are going to use method of calculating average license price is to not use sold counts but actually just listed prices on the pricing page so: +1=159.00 = 159 per user +5=695.00=139 per user +10=1190=119 per user +15=1650=110 per user +20=1980 =99 per user +Average here = 125.00 per user so not much difference +This means that the average v7 subscription charge per year is $32 +Which means that the average v7 true initial license charge is actually $93 (93 + 35%=125) +And this also means that in v7 the subscription is worth $2.66 (avg) per month (32 / 12) +o if v7 was using our new system it would mean we would charge $93 for a license plus $2.66 for the initial 30 days support for a total of $95 (rounded down), then if they purchase a sub it would (in v7 prices) be $32 for 12 months. + +Meaning if use 25% inflation figure v8 would be 95+25% $118 for the license and $40 license for a years sub and updates. + +##### How does it work out? + +Gamed out new pricing based on above system + +All license prices come with 30 days support and updates (maintenance) + +After 30 days runs out they can NOT upgrade nor get support without a subscription. + +**TODO** Some kind of system to prevent just not buying a sub until they need it like a discount to maintain one but a higher charge to purchase, i.e. it's like they are buying a new release in addition to the subscription + +A yearly optional maintenance subscription can be bought which entitles to support and updates including major version releases, if it lapses there is some kind of extra high penalty to start up again but it's not full price so there is _something_ of value in being a prior customer (or is that being too nice?) + +V7 PRICING FOR COMPARISON +(price/maintenance WHICH FOR v7 is only in _SUBSEQUENT_ years, not first year) +1 user $159/$55.65 +5 user $695/$243 (139/48.6 each) +10 user $1190/$416 (119 each) +15 user $1650/$577 (110 each) +20 user $1980/$693 (99 each) +50 user $3950/$1382 (79 each) +QBI $99/$34.65 + +HYPOTHETICAL PRICING @ 35% maintenance +1 user = $118, maint=$40 ($158 if bought together) +5 user = $590, maint=$200 ($790 together) +10 user = $1180, maint=$400 ($1580 together) +15 user = $1770, maint=$600 ($2370 together) +20 user = $2360, maint=$800 ($3160 together) +50 user = $5900, maint=$2000 ($7900 together) + +Ok, that's too low almost no one is paying more, I still like the flat license price for this though, it's easier than the fuckery out there in the world and these are really crazy value for money, like 100:1 or more so let's revamp, the cost of development is uber high so lets bump way up the subscriptions, instead of 35% let's make it 50% and see + +HYPOTHETICAL PRICING @ 50% maintenance +1 user = $118, maint=$59 ($177 if bought together) (minimum 2?? this is way less than a single hour bill out rate) +5 user = $590, maint=$295 ($885 together) +10 user = $1180, maint=$590 ($1770 together) (Here and down are our bread and butter and they're barely higher) +15 user = $1770, maint=$885 ($2655 together) +20 user = $2360, maint=$1180 ($3540 together) (70% increase in yearly maint) +50 user = $5900, maint=$2950 ($8850 together) + +SCENARIO: KEEP V7 LEVELS, INCREASE EACH LEVEL BY 25% INFLATION AND MAINT %50 INSTEAD OF 35% +(price/maintenance) +1 user $200/$100 +5 user $870/$435 +10 user $1485/$742 +15 user $2060/$1031 +20 user $2475/$1237 +50 user $4937/$2468 +QBI $125/$63 + +SCENARIO: with 30 days support only and optional yearl maint. A tech is worth on average 100 / hr so do maint at 100 and initial at 35% higher flat no discounts +\*\*\* MARKET THIS AS $ per tech per month as it's fucking cheap when you look at it that way + +POTENTIAL **UPDATE, ACTUAL** NEW PRICING +(UP FRONT PRICE / YEARLY SUPPORT AND UPDATES) +1 user $135/$100 +5 user $675/$500 +10 user $1350/$1000 +15 user $2025/$1500 +20 user $2700/$2000 +50 user $6750/$5000 +QBI $135/$100 + +So hypothetical profit if using this system with 2022 subscription counts would be: +340 licenses \* 100 dollars = 34,000 dollars +QBI, 10 active @100=$1000 so 35,000 dollars per year + +Newcomers would double that in the first year becuase they buy the license then the main or maybe not maint but we get more than the maint in year one anyway. + +#### AVERAGE HOURLY BILILING RATES FOR SERVICE COMPANIES 2022 + +Plumber 75-150 average 112.50 +Electrician low average 75/hr +General contractor 45 +Auto Mechanic 102 +HVAC 150 +96.8/hr average so let's say 100 / hr + +#### SUBSCRIPTION / SAAS + +So it's clear looking over things that there's a cost for a droplet and adding x more users doesn't really add to that for us so if a $5 droplet works for a single user up to let's say at least 10 users, just speculation, we make way more money off the 10 user. The price for a single user must be increased to accomodate that or...we host more than one site on a single droplet to offset the cost, but charge enough to move them up to a higher level without needing to increase the charges to compensate. + +so pricing should be the highest normal droplet we can use + +The range is $64/month/user at the top end (minimum 2 users though) to around 12-15/month/user at the lower end which usually corresponds to 10+ user counts +Single user is always most expensive for most sites and I can understand why as it's economy of scale + +Current subs revenue per TECH license is 3.50 per month ($4.375 in 2022 dollars) and we have 340 technician licenses active right now. If they all paid monthly that's $1190 dollars a month in current pricing in support and updates subscriptions alone. In 2022 dollars with inflation that would be 1,487.00 which is only $17,850 yearly revenue, no wonder we're poor. This is unsustainable clearly. + +So let's imagine a different scenario, let's say there are .75 users for every service tech in a business so 340 + 255 = 595 users if they all switched to SAAS + +WAYS TO FIGURE + +WHAT IF we just charge double the perpetual maint rate? + +Maybe double the perpetual maint price so 16.66 rounded up to 17 per month per user? + +WHAT IF we just charge $35/mo (yearly 12.5% discount and the assumed price everyone would pay) $40/mo (monthly) about triple the maint price + +so for one user our profit if they paid yearly would be assuming a 7 dollar server 35-7=336 dollars per year, maybe not worth it so minimum 2?? +two user 70/mo-7=63\*12=756/year, that should be less than 1 hour of my time average per month to be worthwhile, like seconds at most + +5 user would be 175/mo-7=168\*12=2000/year + +10 user would be 350/mo-7\*12=4116/year + +So going by the wild assed assumption that our current subscriber levels all went to SAAS the profit could be: +1 user->2user so assume we keep only half of them, there are 30 singles so 15*70=1050-15*7\*12= + +###### Server costs + +Once we're rolling with cash flow it would make sense to just rent a huge dedicated server and put multiple accounts on it as it would be very controlled we'd know who and what is using it not shared etc + +Ideally we'd put users individually on the cheapest server available but in practice maybe what we want to do is rent a bigger dedicated server and split users on it (though they need to be in the same datacenter for that to work) + +We'll allocate the $7/mo server to single users so that's 30 techs plus .75 staff so 52.5 users total +BUT, let's count on the 14 a month server and worst case the 68/month server dedicated +We can probably allocate the same 7/mo to all others and maybe incrementally go up a few before hitting the Absolute worst case scenario i.e. most expensive server ever likely required (not counting storage extra or email costs if we provide which maybe is an add-on??) would likely ever be is dedicated 8gb 2 cpu amnd 50gb storage at $68.00 per month so that would likely be the top end requirement for a pretty big user. + +Maybe it's an idea to always use NAS block storage out of the gate for attachments and just expect it or add-on 10bucks for 100gb block storage is an option. +What is scotts total storage again I forget I think it was 4gb or something?? + +So the cheapest license cost if we said 12 per user for 10 and up hypothetically would mean a worst case scenario of 120 bucks a month revenu for a 21 dollar a month outlay for server so profit of about 100 bucks a month on the 10 users SAAS versus current revenu of 35 dollars a month for those same 10 users hmm... + +of course we would use the cheapest server set up monitoring and bump them up if we see it pegging out, going to take some time to figure this shit out + +Maybe the way to look at this is a basic price with options that directly correspond to digitalocean add-on's so fixed to 25gb total storage for the droplet initially + +##### Add-ons upgrades + +- STORAGE: So we assume a cheap ass droplet and offer add-on's like more storage attachment space for XX / mo which is really just the digitalocean pricing plus a premium overhead for contingencies if they want it +- RAM / CPU: maybe we offer basic server but they can upgrade to levels for additional per month +- BACKUP: self is free vs we do is whatever digitalocean charges plus a premium +- Email, if they don't bring their own account we provide at a cost / sendgrid or whatever (I suspect people expect this to be included if turnkey from a phone or something) +- Anything we get billed extra for, as it stands we don't care how powerful they need, ayanova supports it, but we don't pay for it, they do. +- So, maybe the way to go is base everything off a tiny or next to tiny server and anything else they can pay for as an upgrade. + ### Comparatives (as of 2022-08-11) @@ -142,207 +381,6 @@ Bottom line $64.00 / user / month So 20% discount for yearly billing instead of monthly seems pretty common -### License types and programs we will offer - -TWO types makes the most sense after considering options: - -#### Perpetual - -- Most similar to current v7 but not exactly the same, subscription is broken out separately -- One time fee, user can use indefinitely -- self installed, hosted and maintained by customer -- least profitable for us long term if they don't buy a maint. subscription -- Without maintenance subscription, eligable for Minor updates only to fix bugs no new features so in other words they buy 8.0.2 they can upgrade to any 8.0.X version release, but not 8.1 as it will be new features added that don't break backward compatibility -- one-time payment, along with the option of a yearly maintenance fee. -- This is basically our current model but we allow upgrades for subscribers -- **HAS CODE IMPLICATIONS** upgrades need to check if allowed based on version number if no maintenance subscription _not_ on date of build. -- TBD - MAJOR RELEASES: not sure yet; if they have an active maintenance subscription can they can upgrade to any newer major releases?? -- If they lapse the subscription they pay a much higher premium but not quite as high as a completely new purchase to reward past loyalty (i.e. maybe 75% or some other % as appropriate of a new purchase or something?) - -#### Subscription license - -- SAAS, pay month to month to keep using it -- We host, maintain and always upgrade they don't need to do a thing but use it -- The subscription model involves recurring payments, typically monthly or yearly. The subscription model can be thought of as “renting” the product instead of “owning” it under the perpetual model. - -#### SWITCHING - -We need a formal declaration of how to switch from one to the other models. - -##### Perpetual to subscription - - they start with perpetual but go fuck it we don't want the hassle anymore so....? - I'm guessing it's just switching to month to month plans and pricing, nothing special except maybe we offer to move their data for them - -##### Subscription to perpetual - - They must purchase a perpetual license as the SAAS license wont' work with perpetual but they can restore their database from the subscription. - -### Pricing for v8 - -#### PERPETUAL - -### Actual price (work below) - -**US$135.00 / tech up front perpetual license** -No volume discounts -Gets you 30 days support and updates and that's it - -**US$100/tech/year maintenance** -No volume discounts -Gets you one year of support and updates to any release - -**UPDATE INFLATION IN TRUE TERMS CLOSE TO 25% so using that** - -IDEA: figure out average license cost over all sizes of existing license sales, factor _that_ with the 25% true inflation and make that the average license price?? - -**update: This concept ok, numbers not exactly what we're going for though** -For active subscribers the average price paid for licenses was 121.25 so with inflation of 19.4 that would be 144.78. -Tech counts by level is single=30 $4,770.00 total sales, 5=85 $11,815.00 total sales, 10=110 $13,090.00 total sales, 15=15 $1650.00 total sales, 20=100 $9,900.00 total sales -(more 10 level techs than any other level and single is kind of lame) -So, if we flat price no volume discount and want the inflation increase, if we stuck with 189 for everyone using the single price that's actually a huge amount for the up to 20 crowd - -The real revenue is in the subscription so a lower-ish license cost is where it's at -But what if no one buys a subscription?? Maybe though, lower up front and no mandatory subscription which must drive some people away, means more sales when you have a sellable product like early days of AyaNova - -This is all about the initial license purchase, not the subscription which I haven't got to yet. Anyway, we were talking about a flat price per license no matter how many you buy. I still like that idea and was doing some calculations to figure it out. If we went that route then the pricing would make more sense to be calculated on the _average_ price per license sold, not on the single user price which would end up being 90% higher for a 20 user site. I used our existing active subscribers as a basis to calculate the average price of each license sold. I totaled the amount we charged for all active licenses then divided by number of techs overall that the represent and came to $121.25 is the average price we charge per license. So with inflation using your calculation of 19.41% that's 144.78 flat rate price per tech. - -Except the current v7 license comes with 1 year of support and updates and I'm thinking of 30 days initial support and updates so basically we're taking away 11 months worth of support and updates which is not nothing either, not sure if that needs to be factored in or not? -The "value" of a support and updates subscription in v7 terms is 35% of initial license price so using the average that's $42.43 a year per tech or $3.40 / month per tech. -So of that 121.25 we are saying that the tech / license portion is 78.82 and the support and updates portion is 42.43 - -**UPDATE CHANGE OF CALCS HERE, PRIOR WAS BASED ON SALES HERE ON OUT WE USE ACTUAL LISTED PRICES** - -Revenue for current v7 users based on subscription only, not new sales and only AyaNova licenses: -Single, 30 active @$55.65 per=$1669 -Up to 5, 17 active @$243=$4131 -Up to 10, 11 active @$416=$4576 -Up to 15, 1 active @$577=$577 -Up to 20, 5 active @$693=$3465 -QBI, 10 active @34.65=$346.00 -So currently $14,418 total revenue for tech subs alone, that needs to like triple just to be worthwhile - -Actual we are going to use method of calculating average license price is to not use sold counts but actually just listed prices on the pricing page so: -1=159.00 = 159 per user -5=695.00=139 per user -10=1190=119 per user -15=1650=110 per user -20=1980 =99 per user -Average here = 125.00 per user so not much difference -This means that the average v7 subscription charge per year is $32 -Which means that the average v7 true initial license charge is actually $93 (93 + 35%=125) -And this also means that in v7 the subscription is worth $2.66 (avg) per month (32 / 12) -o if v7 was using our new system it would mean we would charge $93 for a license plus $2.66 for the initial 30 days support for a total of $95 (rounded down), then if they purchase a sub it would (in v7 prices) be $32 for 12 months. - -Meaning if use 25% inflation figure v8 would be 95+25% $118 for the license and $40 license for a years sub and updates. - -##### How does it work out? - -Gamed out new pricing based on above system - -All license prices come with 30 days support and updates (maintenance) - -After 30 days runs out they can NOT upgrade nor get support without a subscription. - -**TODO** Some kind of system to prevent just not buying a sub until they need it like a discount to maintain one but a higher charge to purchase, i.e. it's like they are buying a new release in addition to the subscription - -A yearly optional maintenance subscription can be bought which entitles to support and updates including major version releases, if it lapses there is some kind of extra high penalty to start up again but it's not full price so there is _something_ of value in being a prior customer (or is that being too nice?) - -V7 PRICING FOR COMPARISON -(price/maintenance) -1 user $159/$55.65 -5 user $695/$243 (139/48.6 each) -10 user $1190/$416 (119 each) -15 user $1650/$577 (110 each) -20 user $1980/$693 (99 each) -50 user $3950/$1382 (79 each) -QBI $99/$34.65 - -HYPOTHETICAL PRICING @ 35% maintenance -1 user = $118, maint=$40 ($158 if bought together) -5 user = $590, maint=$200 ($790 together) -10 user = $1180, maint=$400 ($1580 together) -15 user = $1770, maint=$600 ($2370 together) -20 user = $2360, maint=$800 ($3160 together) -50 user = $5900, maint=$2000 ($7900 together) - -Ok, that's too low almost no one is paying more, I still like the flat license price for this though, it's easier than the fuckery out there in the world and these are really crazy value for money, like 100:1 or more so let's revamp, the cost of development is uber high so lets bump way up the subscriptions, instead of 35% let's make it 50% and see - -HYPOTHETICAL PRICING @ 50% maintenance -1 user = $118, maint=$59 ($177 if bought together) (minimum 2?? this is way less than a single hour bill out rate) -5 user = $590, maint=$295 ($885 together) -10 user = $1180, maint=$590 ($1770 together) (Here and down are our bread and butter and they're barely higher) -15 user = $1770, maint=$885 ($2655 together) -20 user = $2360, maint=$1180 ($3540 together) (70% increase in yearly maint) -50 user = $5900, maint=$2950 ($8850 together) - -SCENARIO: KEEP V7 LEVELS, INCREASE EACH LEVEL BY 25% INFLATION AND MAINT %50 INSTEAD OF 35% -(price/maintenance) -1 user $200/$100 -5 user $870/$435 -10 user $1485/$742 -15 user $2060/$1031 -20 user $2475/$1237 -50 user $4937/$2468 -QBI $125/$63 - -SCENARIO: A tech is worth on average 100 / hr so do maint at 100 and initial at 35% higher flat no discounts -(price/maintenance) -1 user $135/$100 -5 user $675/$500 -10 user $1350/$1000 -15 user $2025/$1500 -20 user $2700/$2000 -50 user $6750/$5000 -QBI $135/$100 - -So hypothetical profit if using this system with 2022 subscription counts would be: -340 licenses \* 100 dollars = 34,000 dollars -QBI, 10 active @100=$1000 so 35,000 dollars per year - -#### AVERAGE HOURLY BILILING RATES FOR SERVICE COMPANIES 2022 - -Plumber 75-150 average 112.50 -Electrician low average 75/hr -General contractor 45 -Auto Mechanic 102 -HVAC 150 -96.8/hr average so let's say 100 / hr - -#### SUBSCRIPTION / SAAS - -So it's clear looking over things that there's a cost for a droplet and adding x more users doesn't really add to that for us so if a $5 droplet works for a single user up to let's say at least 10 users, just speculation, we make way more money off the 10 user. The price for a single user must be increased to accomodate that or...we host more than one site on a single droplet to offset the cost, but charge enough to move them up to a higher level without needing to increase the charges to compensate. - -so pricing should be the highest normal droplet we can use - -The range is $64/month/user at the top end (minimum 2 users though) to around 12-15/month/user at the lower end which usually corresponds to 10+ user counts -Single user is always most expensive for most sites and I can understand why as it's economy of scale - -Current subs revenue per TECH license is 3.50 per month ($4.375 in 2022 dollars) and we have 340 technician licenses active right now. If they all paid monthly that's $1190 dollars a month in current pricing in support and updates subscriptions alone. In 2022 dollars with inflation that would be 1,487.00 which is only $17,850 yearly revenue, no wonder we're poor. This is unsustainable clearly. - -So let's imagine a different scenario, let's say there are .75 users for every service tech in a business so 340 + 255 = 595 users if they all switched to SAAS - -###### Server costs - -We'll allocate the $7/mo server to single users so that's 30 techs plus .75 staff so 52.5 users total -We can probably allocate the same 7/mo to all others and maybe incrementally go up a few before hitting the Absolute worst case scenario i.e. most expensive server ever likely required (not counting storage extra or email costs if we provide which maybe is an add-on??) would likely ever be is dedicated 8gb 2 cpu amnd 50gb storage at $68.00 per month so that would likely be the top end requirement for a pretty big user. - -Maybe it's an idea to always use NAS block storage out of the gate for attachments and just expect it or add-on 10bucks for 100gb block storage is an option. -What is scotts total storage again I forget I think it was 4gb or something?? - -So the cheapest license cost if we said 12 per user for 10 and up hypothetically would mean a worst case scenario of 120 bucks a month revenu for a 21 dollar a month outlay for server so profit of about 100 bucks a month on the 10 users SAAS versus current revenu of 35 dollars a month for those same 10 users hmm... - -of course we would use the cheapest server set up monitoring and bump them up if we see it pegging out, going to take some time to figure this shit out - -Maybe the way to look at this is a basic price with options that directly correspond to digitalocean add-on's so fixed to 25gb total storage for the droplet initially - -##### Add-ons upgrades - -- STORAGE: So we assume a cheap ass droplet and offer add-on's like more storage attachment space for XX / mo which is really just the digitalocean pricing plus a premium overhead for contingencies if they want it -- RAM / CPU: maybe we offer basic server but they can upgrade to levels for additional per month -- BACKUP: self is free vs we do is whatever digitalocean charges plus a premium -- Anything we get billed extra for, as it stands we don't care how powerful they need, ayanova supports it, but we don't pay for it, they do. - #### New idea perpetual vs subscription There is something to be said for a subscription model even if we don't host it. In this system a "perpetual license" is a one time purchase and comes with no support or maybe a very short window they get to use it forever afterwards. The subscription is a rental service only aka a "term license", if they stop paying it stops working. Separate from hosted AyaNova. diff --git a/devdocs/todo.txt b/devdocs/todo.txt index c7e06422..d16e3af2 100644 --- a/devdocs/todo.txt +++ b/devdocs/todo.txt @@ -1,18 +1,39 @@ # now +todo: TEST digitalocean, set up a server in every data center far away from here, test ayanvoa on each one and see how peformant it is perceptually, i.e. does it even matter which server we're locating people on?? +todo: cheapest storage for attachments possible, test out creating a new droplet with block storage to hold the attachment files to see if it's possible, how it's done and mapped etc + +todo: metrics are fucked up for some reason + in fact I'd just like to reliably know how much disk space is in use etc, like the size of the db versus attachments in simple text + +todo: on ops server info page show main disk usage of db, attachments free memory in plain text unlike the metrics which are fucking useless + +todo: v8migrate Contracts in v7 were allowed to have no name, add code to invent a name on migration so it doesn't bomb + todo: QBI don't check build date against expiry date if it's a trial, i.e. they can upgrade a trial all they want as long as it's not expired key todo: License agreement changes for SAAS customers, this is huge and I hadn't thought of it before, wording is very different in this case, fuck, need to wing it I guess https://flgpartners.com/saas-pivots-transitions-perpetual-to-subscription-saas-models/ needs to be available separately in docs Needs a disclaimer about the right to refuse service if becomes too high usage i.e. subject to usage limitations to cover our ass and check with D.O. how to clamp that from going into expensive territory for us? + We also need to mention we will size their servers to adequate work load or whatever meaning we will put them on the smallest cost server that will sustain their needs. todo: ROCKFISH different license key signature if it's a SAAS license this solves a lot of problems, effectively it's unlicensed if they move their data to their own hardware without a key requires that AyaNova have different internal key to check, maybe it's actually a build switch for that since only we will be running our SAAS version. +todo: ROCKFISH can't make a trial key for users in the UI, only when it's requested from ayanova. + I'm really not sure if this is an issue or not but putting it on the list in case + +todo: DOCS - remove all mention of digitalocean from the install and other docs since we're going to be using them for SAAS we don't want to make it simple for people to host themselves, just say Linux and leave it at that. + + +todo: RENTAL terminology needs to change to "SUBSCRIPTION" in rockfish and in client and at server, RENT should not appear anywhere, it's awkward and old fashioned +todo: SUBSCRIPTION should only refer to month to month SAAS feature / plan. Support and upgrades should be referred to as Maintenance or Support and upgrades without the word "subscription" +TODO: above changes at SHAREIT with the product descriptions as well so no confusion + todo: if SAAS needs to do some things differently: if it helps, consider a different build flag since SAAS is like an internal build for us SERVER - count total active users towards license not techs only @@ -34,7 +55,8 @@ todo: if SAAS needs to do some things differently: maybe it's a plust because we can say you get a maximum storage for attachments of XXGb based on droplet size and it's an add-on to go with a bigger server?? Because some people need it and will pay for it if it's offered. SERVER - how to cap bandwidth, storage, whatever we get billed extra for from d.o. - + CLIENT - NEEDs to report that a new version was installed to the user with link to changes when we publish a new version so they can be made aware + EMAIL - research how to send emails with a reliable provider i.e. sendgrid or alike and what the costs and add-on price for that would be. todo: each droplet for SAAS would need it's own ssh key and password to avoid eggs in one basket security scenario also would need to securely store the passwords somewhere really really fucking secure like offline in a special keypass only accessible from certain hardware maybe with a digital key or something I dont' know @@ -68,7 +90,7 @@ todo: SAAS different license checking method to ensure only *we* are hosting it -todo: test out creating a new droplet with block storage to hold the attachment files to see if it's possible, how it's done and mapped etc + todo: test out automatic backups using digitalocean's built in system with their pricing and how that would work with our backup system for the db etc @@ -180,6 +202,7 @@ todo: validate the rate limiting by tailing the error log and hitting test raven todo: QBOI deprecation BUT...this is probably going to be needed for future again so keep on back burner + YES: EVERYONE OF OUR COMPETITORS THAT HAS ANY QB HAS QBONLINE it's the desktop that is rarer done: QBOI removed from website except still in docs and in trial installer done: Product is set to inactive with ShareIt, no one can buy it now OUTSTANDING: I will need to remove it from the AyaNova installer completely diff --git a/docs/8.0/ayanova/docs/ops-install-linux-server.md b/docs/8.0/ayanova/docs/ops-install-linux-server.md index 007b40a1..b61f8ab8 100644 --- a/docs/8.0/ayanova/docs/ops-install-linux-server.md +++ b/docs/8.0/ayanova/docs/ops-install-linux-server.md @@ -18,7 +18,7 @@ These instructions were compiled and tested on an Ubuntu 20 server OS however an ## Hosted linux -There is no significant difference in setup between a linux server physically on premise or hosted from a hosting provider such as [Digitalocean](https://www.digitalocean.com/) or [Linode](https://www.linode.com/) with SSH shell access. +There is no significant difference in setup between a linux server physically on premise or hosted from a hosting provider with SSH shell access. The largest demand put on an AyaNova server is report rendering so if it will be common to render reports with dozens or hundreds of pages you will find that more memory and CPUs will be beneficial. @@ -51,7 +51,7 @@ If your server has 2 GB of RAM or less or will be rendering many reports simulta Swap memory should be considered a stopgap emergency measure only and not relied upon regularly to ensure sufficient memory. If the swap file is being accessed regularly it's a clear indication that more ram needs to be installed or allocated. -DigitalOcean [has a guide](https://www.digitalocean.com/community/tutorials/how-to-add-swap-space-on-ubuntu-20-04) for Ubuntu droplets that walks through the process or see your distribution documentation for setting this. We recommend double the size of ram allocated to swap space so if you have 1GB of ram then a 2GB swap space is appropriate. +We recommend double the size of ram allocated to swap space so if you have 1GB of ram then a 2GB swap space is appropriate. ### 2\. Install ASP.NET Core runtime @@ -296,7 +296,7 @@ The following items will be required to configure AyaNova on a linux server for ##### Static IP address -A static IP address is **required** to host AyaNova on the internet. If you have a virtual or physical server through a server hosting provider such as Digitalocean or Linode then you will already have a static IP address. If you are hosting your own server on premise on a private office network connecting to the internet via an Internet Service Provider you will need to ensure your ISP has provided you with a static IP address and if they haven't you will need to request a static IP address and have that address on hand before proceeding. +A static IP address is **required** to host AyaNova on the internet. If you have a virtual or physical server through a server hosting provider then you will already have a static IP address. If you are hosting your own server on premise on a private office network connecting to the internet via an Internet Service Provider you will need to ensure your ISP has provided you with a static IP address and if they haven't you will need to request a static IP address and have that address on hand before proceeding. ##### Registered domain name diff --git a/docs/8.0/ayanova/docs/ops-install-linux.md b/docs/8.0/ayanova/docs/ops-install-linux.md index 0fbaf3cf..fc3e9955 100644 --- a/docs/8.0/ayanova/docs/ops-install-linux.md +++ b/docs/8.0/ayanova/docs/ops-install-linux.md @@ -10,6 +10,6 @@ This type of install is appropriate for using AyaNova locally on a single person ## Shared multiple user server installation -When more than one user will be accessing AyaNova it can be set up on a Linux server either on-premise or with a Linux hosting provider, for example [Digitalocean](https://www.digitalocean.com/) or [Linode](https://www.linode.com/) that provides full shell access to a virtual or physical server. +When more than one user will be accessing AyaNova it can be set up on a Linux server either on-premise or with a Linux hosting provider, that provides full shell access to a virtual or physical server. [Server install guide](ops-install-linux-server.md) diff --git a/docs/8.0/ayanova/docs/ops-intro.md b/docs/8.0/ayanova/docs/ops-intro.md index b85301ba..145c345d 100644 --- a/docs/8.0/ayanova/docs/ops-intro.md +++ b/docs/8.0/ayanova/docs/ops-intro.md @@ -26,7 +26,7 @@ The "front end" of AyaNova also known as the "client" is the user interface for AyaNova supports many configurations from single user on a single computer, multiple office users on a local area network to large corporate scenarios where AyaNova is accessed globally from the internet. -AyaNova server can be installed on many platforms including a single computer for private use, on-premise with a office network server or hosted on a 3rd party internet hosting provider such as [DigitalOcean](https://www.digitalocean.com/). +AyaNova server can be installed on many platforms including a single computer for private use, on-premise with a office network server or hosted on a 3rd party internet hosting provider. The [installation guide](ops-install.md) is the starting point for installing the AyaNova server to meet specific needs.