Create an excel or google sheet diagram for the calculation of variable costs -- 2
Budget: $2 – $8 USD
I aim to create a SaaS (Software as a Service) platform. This service will allow users to register their own voice and then upload real voice messages. These messages can be converted by AI into various languages and returned as AI-generated voice messages in their original voice. Users can use this feature for fun, like trolling their friends, or as a translator during travels.
The resources required for this SaaS are:
1. Stripe as a payment provider: 1.5% + €0.25 per transaction.
2. Zapier Plan: €69 for 2,000 tasks/month, €125 for 5,000 tasks/month, €182 for 10,000 tasks/month.
3. For speech-to-text conversion and text-to-text translation: 3.1 GPT-3.5 via API, €0.002/1,000 tokens for output, €0.001/1,000 tokens for input. 3.2 GPT-4.5 via API, €0.03/1,000 tokens for output, €0.01/1,000 tokens for input. It is assumed that both models will be used equally.
4. Chatbot software: $15/month for up to 1,000 users, an additional $5/month for every additional 1,000 users up to 5,000 in total, then $20/month for every additional 5,000 users.
5. Elevenlabs for text-to-speech conversion in custom voices:
• $22/month for up to 100,000 converted characters, a maximum of 30 voices, and $0.30/1,000 characters beyond 100,000.
• $99/month for 31-160 voices, up to 500,000 converted characters in total, and $0.24/1,000 characters beyond 500,000.
• $330/month for 161-660 voices and 2,000,000 converted characters in total, and $0.18/1,000 characters beyond 2,000,000. For more, an enterprise solution needs to be requested from Elevenlabs.
Please calculate the total expenses per customer for the first, second, and third months after launch. Provide a diagram with € on the Y-axis and the number of customers on the X-axis in 5-step increments (5, 10, 15… 600) connecting the 132 points for each of the three months (660 divided by 5) with separate lines.. The lines should be distinct in color. Also, under the curve for the first month, show the integrals of the individual cost factors in different shades, with each shade representing a cost point, cumulatively building up to the total cost line. Then the same in another diagram from 100-10.000 customers in 100-step increments.
The following assumptions are made:
1. Each customer uses 3 Zapier operations for subscription completion. If a payment fails (once a month), it requires another 3 operations, happening in 10% of all contracts. An additional 3 operations are needed for a cancellation.
2. It is assumed that 60% of all contracts are canceled after the first subscription.
3. The maximum number of characters for conversion is shared among customers. For example, 160 users share 500,000 characters, and 165 users share 2,000,000 characters. For every fifth customer calculated on the X-axis, the costs for OpenAI differ. It is assumed that 4 characters equal 1 token.
4. Each user is expected to receive a total of 10 minutes of voice message playback per month. According to the Pareto principle: 80% of users only use 20% of their limit, 19% use 80%, and 1% use 100%. Calculate the expected value of quota consumption as a fixed number for the calculation of token/character-based costs for OpenAI and Elevenlabs. Assume a plausible conversion rate from playback duration to characters or tokens.
5. It is assumed that the sum of input and output tokens for OpenAI is exactly equal.
If further assumptions are needed, please let me know or make plausible assumptions yourself.
each 660 users cost additional 330 USD in elevenlabs.
so it's linear, without an enterprise agreement with elevenlabs.
The resources required for this SaaS are:
1. Stripe as a payment provider: 1.5% + €0.25 per transaction.
2. Zapier Plan: €69 for 2,000 tasks/month, €125 for 5,000 tasks/month, €182 for 10,000 tasks/month.
3. For speech-to-text conversion and text-to-text translation: 3.1 GPT-3.5 via API, €0.002/1,000 tokens for output, €0.001/1,000 tokens for input. 3.2 GPT-4.5 via API, €0.03/1,000 tokens for output, €0.01/1,000 tokens for input. It is assumed that both models will be used equally.
4. Chatbot software: $15/month for up to 1,000 users, an additional $5/month for every additional 1,000 users up to 5,000 in total, then $20/month for every additional 5,000 users.
5. Elevenlabs for text-to-speech conversion in custom voices:
• $22/month for up to 100,000 converted characters, a maximum of 30 voices, and $0.30/1,000 characters beyond 100,000.
• $99/month for 31-160 voices, up to 500,000 converted characters in total, and $0.24/1,000 characters beyond 500,000.
• $330/month for 161-660 voices and 2,000,000 converted characters in total, and $0.18/1,000 characters beyond 2,000,000. For more, an enterprise solution needs to be requested from Elevenlabs.
Please calculate the total expenses per customer for the first, second, and third months after launch. Provide a diagram with € on the Y-axis and the number of customers on the X-axis in 5-step increments (5, 10, 15… 600) connecting the 132 points for each of the three months (660 divided by 5) with separate lines.. The lines should be distinct in color. Also, under the curve for the first month, show the integrals of the individual cost factors in different shades, with each shade representing a cost point, cumulatively building up to the total cost line. Then the same in another diagram from 100-10.000 customers in 100-step increments.
The following assumptions are made:
1. Each customer uses 3 Zapier operations for subscription completion. If a payment fails (once a month), it requires another 3 operations, happening in 10% of all contracts. An additional 3 operations are needed for a cancellation.
2. It is assumed that 60% of all contracts are canceled after the first subscription.
3. The maximum number of characters for conversion is shared among customers. For example, 160 users share 500,000 characters, and 165 users share 2,000,000 characters. For every fifth customer calculated on the X-axis, the costs for OpenAI differ. It is assumed that 4 characters equal 1 token.
4. Each user is expected to receive a total of 10 minutes of voice message playback per month. According to the Pareto principle: 80% of users only use 20% of their limit, 19% use 80%, and 1% use 100%. Calculate the expected value of quota consumption as a fixed number for the calculation of token/character-based costs for OpenAI and Elevenlabs. Assume a plausible conversion rate from playback duration to characters or tokens.
5. It is assumed that the sum of input and output tokens for OpenAI is exactly equal.
If further assumptions are needed, please let me know or make plausible assumptions yourself.
each 660 users cost additional 330 USD in elevenlabs.
so it's linear, without an enterprise agreement with elevenlabs.