Mocke - Mock email campaigns: know your reply rate without launching
Mocke AI agent mocks running email campaigns. Get back results in 1 min, such as open, reply or unsubscribe. Mocke also reveals why each lead didn't open emails, reported spam or ignored. Unlock insights of campaign failure (or success!) not seen in real life.
Replies
Mocke
Hey AI enthusiasts,
It's @thisismattsun and @daniel_wasserlauf from Mocke (YC S22). We are beyond excited to launch a product that's never seen before with the Product Hunt community! To get you started, I'm going to share the problem we are trying to solve for you, describe the product itself, and talk a bit about ourselves at the end. Let's get going 🚅🚅🚅🚅🚅🚅
The problem, that we take personally 😕
At Mocke, we send a huge amount of cold emails (for our startup journey, see it in the end). Our daily cold email sending volume is up to 1,000,000 per day to the entire world. And we've seen a wild range of performance outcomes: the best around 80% open / 25% reply / 10% interest, and the worst would be 0 interest or replies. Granted, cold email is hard because of all factors driving the results: deliverability, messaging, cta, brand awareness, competition, PMF, etc. At our scale, we've built possibly all tools and AI to maximize the performance, except one thing:
On one day, we've seen 10% interest rate for a given product, which easily generates 100 leads per day at our scale. On the next day, for the same product, we tried different messaging, just a minor tweak, and prospect list, we've got 0 replies. (rest assured, deliverability isn't our problem, our open rate never drops below 65% when we track it).
The problem is messaging-market fit. We cannot tell whether a messaging is going to resonate with this list of leads/persona/decision maker until we test it. We have to test it.
So here are the three core problems we are struggling with, and I'm sure you do as well:
💔 SLOW - Cold email takes a long time to test. Every campaign with multi-step sequence takes 2 weeks - 1 month to test.
💔 GUESS GAME - Cold email gives no feedback. If you have a 0% interest rate campaign sent to 1,000 people, why do people not respond? What's the problem? No one tells you, among those 1,000 people, why they don't care about your messaging. So you guess again and spend two weeks to run another email copy.
💔 WASTE - You are burning through real people, real customers. You can't test too hard on these real customers, as it's going to harm your reputation.
This is a real challenge that everyone has to go through. There is no solution so far. Until today.
Meet Mocke
Today we are introducing Mocke, the world's first, and only, AI agent that mocks cold email campaigns exactly like real campaigns. Here is how Mocke works:
📁 Upload your prepared campaign .csv file as usual to Mocke, 1-click to launch the (mocked) campaign
⏰ In as little as 1 minute, the emails are sent, and the campaign results are ready for you
📈 You can view all people's reply, positive or negative, forward, unsubscribe, out of office, etc. like a real campaign
📊 Mocke not only reveals your predicted open rate, reply rate, interest rate and unsubscribe rate, but you also get the following critical information:
🚯 How many people clicked Report SPAM
📨 How many people forwarded internally without letting you know
🗑️ How many people ignore/archive/delete
🤔 The leads' thoughts behind why people opened email, why people read 2nd time, why people reply, etc.
💡 Based on visible campaign outcome, and hidden lead actions above, AI agent analyzes the campaign to provide unseen insights in real campaigns instantly.
With Mocke, the game with cold email has forever changed.
❤️🔥 Instead of waiting for 2 weeks, you get results in 1 min
❤️🔥 Instead of guessing game "why people didn't reply", you know the exact reasons
❤️🔥 Instead of testing against real customers, you test against them mocked, as many times as you wish
Personally, I can't believe there is no such product in the world. We searched, but couldn't find any. So we made it for you, all of you looking at this comment right now.
Our ask
We would love to have everyone to take a look and give it a try! We have a live demo available on website, a free forever plan, and deep discount for this launch, whichever works best for you!
Your support 💪💪💪 means the world to us! Your support helps to spread the word to the entire community so that more people can benefit from this life-changing AI agent!!
------------------------ I'm a humble divider line ------------------------
A bit of background about ourselves. Daniel and I are cofounders. We first met in 2021 (maybe?) in a previous company where we worked together on an infra SaaS product. In 2022, we started to build a business together within the YCombinator S22 batch (yes, same company currently launching Mocke). In the past 3+ years, we've built 3 products, did 1 major pivot, talked to maybe 500 investors, served customers from trillion dollar public companies to 1-people pre-seed startups, etc. I used to code, now I vibe designing, vibe marketing, sometimes vibe prototyping. We also have a team that operates globally across 4 time zones.
Interesting concept with Mocke, but I'm curious how accurately it simulates human decision-making across different industries. Will future versions include integration with actual email platforms to compare mock vs. real campaign results?
Mocke
@sujingshen Hey Sujingshen, that's a really sharp question! We actually haven't thought about integrating with real email platform to do side-to-side comparison, but it's going to be amazing I guess! Thanks for suggestion.
In terms of accuracy, I think the best way to see it is as qualitative way of analysis rather than quantitative results. Mock is mock, at the end of day :-)
Triforce Todos
Mocke
@team @abod_rehman Hey Abdul, you are absolutely right. This is the last missing piece of cold email. Have you seen anything like Mocke before? Just curious!
Ace I love this idea!
Just curious, are the simulated open/reply/interest rates based on historical benchmarks or predictive modeling?
Mocke
@justin2025 That's a really good question! I can write a paper on this haha! Let me first upvote this question.
So our simulation does not use historical benchmark, nor does it use predictive modeling. We set up a real world view for each email recipient, and let the email recipient makes near-real reaction (first reaction, second read, draft reply / take actions).
The world view is the hard part. We simulate about 50 real world dimensions, such as day, job, company, personality, role, energy level, mood, past experience, education, etc.
So I think it's performing better than historical benchmark, because there is no such a data set that's big enough and reliable.
I'm also not aware of a good predictive model either. Is there one you would suggest we looking into more?
Tidyread
This is impressive, a real game changer in the world of email marketing! You truly understand the pain points, and this product is super helpful for analyzing campaign performance and refining our emails. I'm curious though, are the results generated by AI ? How do you ensure the AI responses align with how real people would actually react? @Mocke (YC S22)
Mocke
@nicoleastor Hey Nicole! Great question. Yes, all results are generated by AI, therefore, they are mocked. But we also show more data than real campaigns. Regarding how do we ensure AI responses align with how real people would react, we actually run AI in a large-world model (not large language model!), in this world view, everything is mocked, such as current work, past experience, communication preference per individual. So the outcome, as we've seen, are usually extremely realistic!
This changes everything for A/B testing—no more waiting weeks to see what resonates. Such a clever way to save time and leads!
I'm very curious about how to increase the response rate?
Mocke
@cyrusandrew
Great question! The easiest answer for Mocke is to have great message market fit. We find that the best reply rate responses are when:
1. Leads actually have the problem you're looking to solve
- We help determine this via our "World Simulation".
2. The message is well constructed.
- We surface "hidden variables" that normally we don't have access to, so that we can make better decisions about how to properly message someone. Maybe they will throw away your message if its too long because their workload is too high, or if it sounds too "mass email like" Previously we would need to guess and now we don't with Mocke.
Haimeta
Congrats on the launch! This looks like a highly efficient product. I’d love to know how accurate the simulation results are
Mocke
@jean_xavier
Great question. :)
We've been back testing against other campaigns we've ran and have gotten similar results! We're launching our first "Mocke Tested" campaign so when that completes we're going to compare, but broadly speaking the qualitative insights after a campaign has completed has been really helpful for us, we learned that certain clients don't respond well to certain value props, but do to others, which has been a meaningful change for us.
Congratulations team for the launch.. Good work done
Mocke
@debarghyaroy Thanks! I hope you get a chance to try it, even if just to see how "you" would respond to certain emails. It has me pegged pretty well.
This looks super impressive, I’ve burned way too much time and goodwill waiting 2-3 weeks just to find out a cold email campaign flopped. The ability to get predicted replies plus see hidden actions like forwards or archive/delete in under a minute feels game-changing.
Quick question: how close have you found the mocked results to be compared with real-world outcomes? If it’s even ~70-80% accurate, that would save teams months of iteration. Excited to give Mocke a spin!
Mocke
@ha_anh_nguyen
> how close have you found the mocked results to be compared with real-world outcomes?
Great question,
We've done some extensive back testing and have gotten very close to real world results for our campaigns. 1-2% difference in reply rate. However I think the real value for us has been the qualitative insights feature, which compares across your whole campaign, "What worked well, and what didn't work well?" So if you send out a thousand messages and no one responds, you can more definitively say what the root cause might be. Its been a real help for us.
PopPop AI Vocal Remover
Congrats on your launch. I'm interested in this email campaign agent. How reliable are the results generated by Mocke (YC S22) for my email campaign?
Mocke
@charlenechen_123 Thanks so much, and the agent is interesting to say the least 😅
In terms of reliability, there are a few different kinds so I want to make sure I understand what you mean.
The first kind, is self consistency. Or "When I run the same group of people with the same message does the simulation give me similar results?"
In this case, we definitely do, we detect spam or mass email like messages consistently, and detect success cases consistently the same as well. So Within the simulation (Mocke) its very consistent. We have some variation we inject into the simulation which helps mimic the real world. EG: "Did the lead read the email after a long meeting?" Or did they read it right when they got into work.
The second kind is consistency is with real world campaigns.
We can't guarantee your results in the real world, since we don't control your deliverability infrastructure. (Doesn't matter if your message is great if Gmail doesn't let it hit their inbox). But we have done back testing against our own campaigns, and have similar results. Our main goal here is to help you feel confident with Mocke, that your messaging is correct, and that your leads are correct. So you can focus on all the other hard parts of cold outbound.
Agnes AI
This is really interesting! cant wait to try out on my campaigns - congrats on the launch!
Mocke
@cruise_chen Thank you so much! Feel free to ping Matt or Myself directly if you have any thoughts, issues or questions. Always happy to talk shop.
This is wild, I'm impressed by how you’ve tackled the messaging-market fit problem head-on. It’s one of those things everyone feels but can rarely isolate or fix efficiently.
We’ve been building Escape Velocity AI to help founders structure their thinking early on (audience, offer, pricing, etc.), and I can see Mocke being a killer complement when refining cold outreach, fast feedback loops without real-world risks.
Curious: have you seen any surprising “mock” insights that completely changed how people rewrote their messages?
Mocke
@andreitudor14 Thank you
> Curious: have you seen any surprising “mock” insights that completely changed how people rewrote their messages?
Yes actually. Very sharp question. We have been doing some lead gen work, and in back testing we found some really interesting targeting issues with our of clients who runs an online marketplace, and some of their targets internally saw issues with either their brand reputation, or channel risk that we never thought about before.
AI Jingle Generator
Was it (vibe) coded on Replit?
Mocke
@adsy_me
Thanks for the question!
Technical Co-founder here. We didn't think about using Replit. We do use some Agentic Programming for the Frontend React Web client, mainly GPT-5 through Cursor.
Backend there is some Agent work, but its a bit harder there since the problems are a bit trickier to contain or explain well into a prompt. Let me know if you have any other technical questions or implementation questions! Always happy to talk about software architecture. 😊😊