Q: Dear TinyTalk Team - I hope this meets you well.
I am writing to express some concerns and seek clarifications regarding the tiered offerings you currently have. My feedback is centered around three main areas:
1.) Bring Your Own Key (BYOK) Policy and Bot Character Limits: I am curious about the rationale behind implementing BYOK while still imposing restrictions on bot character limits. The essence of BYOK, from my perspective, is to allow users the flexibility to utilize their resources as needed. Could you explain the decision to limit character counts despite this?
2. Access to GPT-4 in Lower Tiers: I am puzzled as to why access to GPT-4 is not available for Tiers 1 & 2, even when users can bring their own keys to access it. It seems counterintuitive to exclude GPT-4 access for these tiers if the costs are covered by users through BYOK. Could you shed light on the justification for this restriction?
3. Bot Quantity vs. Character Limit in Higher Plans: I am trying to understand the logic behind offering a larger number of bots in higher plans, which are then limited to only half the character limit per bot compared to the cheapest plan. This approach seems to diminish the value of having more bots. Could you explain the reasoning behind this structure?
From an outside perspective, it appears that the most economical and valuable plan is, surprisingly, the cheapest one. While the user interface of your product is appealing, the limitations imposed across various aspects may deter potential users from considering it as a viable alternative to existing chatbot platforms.
I believe re-evaluating the value proposition of your tiered offerings could significantly enhance the attractiveness of your product.
Thank you for your attention to these matters. I wish you the best of luck with your launch and look forward to your response.
Best regards!
tinytalk
May 15, 2024A: Hi! 👋
Thank you very much for the questions, feedback and your openness ❤️
I'd like to start by emphasizing that we continuously gather and integrate user feedback to refine our offers and believe that pricing structure should evolve based on actual user needs and market demand.
1. There are two topics here, basically for training characters, the BYOK does not matter that much, the model that is used for training/embedding is very much affordable but it is the cost of vector embeddings that needs to be created, stored, accessed and backed up. The bigger character limit a user has, the more space they allocate on the vector database.
The second part of this question is the BYOK, which at some level overlaps with the reasoning above. Even if you do BYOK, as our customer your requests are still passing through our infrastructure and resources. If you happen to have millions of visitors on your website/apps, that potentially means we have to process millions of requests per day/month, store the data points those visitors generate (like chat history, analytics, logs) and accumulate an ever increasing cost for life.
If I have to think of an analogy, first thing comes to mind is a highway toll system. Yes, you may have purchased a car that you own and you likely pay a road tax per year and yet when driving through certain highways you still need to pay tolls. Because those highways need to be maintained to manage traffic flow, road quality, ensure safety and so on, here is no different.
2. Our tiered pricing model is designed with the intention of offering increasing value at each level. As is common in tiered systems, each tier in our structure is crafted to provide more features, higher limits, and broader access to our platform's capabilities. This progression ensures that businesses of all sizes and needs can find a plan that suits them, allowing for scalability and flexibility as those needs change.
3. So this also depends a bit on the business needs, so far we have seen businesses focus on 6-8 chatbots with heavy training while they use the rest for proof of concept, prototyping, ideation or smaller tasks. We don't limit how you use your training characters per bot, instead you get to choose if you want to go all in on a single bot with all the training characters or distribute it as you wish.
I understand your overall concerns around limits, and if you take a look at our tier 7, which offers unlimited chatbots, does that then dictate unlimited everything.. I believe not and it simply can't if we want to be realistic, because we are offering a life time deal, and we want to keep our promise to provide a service to all Sumo-lings, be here around to get your feedback and build a scaleable, sustainable product where all of our customers can use our resources in a fair way.
If the cheapest plan is the most economical and valuable to you, then that's awesome news, we would be glad to have you on board!
Incase you still have doubts or counter arguments, I'd love to continue our conversation by email over at support@tinytalk.ai If you feel everybody should be able to read our exchange, also feel free to post here.
Having written all that, I want to point out that we will be offering monthly add-on packs to help you increase your limits in a sustainable way, so we get to scale proportionally and when you hit a million active users, you get to serve them smoothly because we are up and running.
Before I round off this message, I want to say one last thing: I'm not fixated on our plan limits, I am all ears/open to feedback, and I want to thank you again for taking the time to post your message 🙏
Hope this helps,
Cheers!