PentoPixis in the Video Creation Tools Industry

PentoPix

pentopix

Blogs are an excellent vehicle for ideas and information, with endless content to read and where you can learn a thing or two. In this slogging thread, our community shares their best suggestions on tech, financial, work, travel, gaming, and pop-culture-themed blogs to follow. This thread occurred in slogging's official #random channel, and has been edited for readability. Slogging Mónica Freitas Jan 5, 2022, 3:26 PM What are some of the best blogs you've found? Tech, financial, work, travel, gaming and pop-culture themed blogs are welcomed🖥 Teri Eyenike Jan 5, 2022, 3:40 PM Work https://www.samjulien.com/ Tech and travel https://petapixel.com/ Abeer Jan 5, 2022, 5:44 PM For tech/dev stuff, I really like https://increment.com/ Jack Boreham Jan 6, 2022, 10:48 AM I have to be honest I don't read many blogs (I know that must sound shocking coming from me haha), any you can recommend, Mónica Freitas. Mónica Freitas Jan 6, 2022, 2:24 PM Teri Eyenike, loved the first one. I immediately got tech from the design. How does he mix tech and travel? Can you share one of your favorite articles? Mónica Freitas Jan 6, 2022, 2:27 PM Abeer, this one is organized by topics, very useful. I saw that there is a focus on programming and financial planning. Is there an article you've found really useful and that you'd like to share? Mónica Freitas Jan 6, 2022, 2:27 PM Jack Boreham, how could you, content writer?! Let me make you a list. Mónica Freitas Jan 6, 2022, 2:30 PM Hubspot is great for topics like web design, marketing, content management, sales and business in general: https://blog.hubspot.com/ Mónica Freitas Jan 6, 2022, 2:31 PM For travel tips, I usually rely on Nomadic Matt: https://www.nomadicmatt.com/ Mónica Freitas Jan 6, 2022, 2:34 PM Inc. is a great resource to keep updated on tech, business and social media trends: https://www.inc.com/technology Mónica Freitas Jan 6, 2022, 2:36 PM PopMatters is great to get info on movies, music, books, etc.: https://www.popmatters.com/ Mónica Freitas Jan 6, 2022, 2:39 PM Gaming-wise, I'd suggest Polygon or The Guardian (not a blog but a good resource): https://www.polygon.com/ Mónica Freitas Jan 6, 2022, 2:39 PM https://www.theguardian.com/technology/gamesblog Jack Boreham Jan 6, 2022, 5:25 PM Mónica Freitas, hahaha, thank you. So what do you look for in a blog? Sara Pinto Jan 7, 2022, 11:21 AM Mónica Freitas, I usually don't follow blogs, but I like one about organization and study. I'm a sucker for planners, so this does it for me haha. But I have to say it's way easier to follow the other social accounts to get the information instead of following blogs. Mónica Freitas Jan 7, 2022, 12:22 PM Jack Boreham, I'm a sucker for aesthetics, so one that appeals to me visually, immediately drags me in. Then, the content: is it clear and well written? Is it useful? If so, you've got a new fan. Mónica Freitas Jan 7, 2022, 12:23 PM Sara Pinto, can you share that one blog you follow? I'm a bit of a procrastinator, so that would be helpful 😂 Sara Pinto Jan 10, 2022, 4:18 PM Mónica Freitas, oh, I forgot to add the link haha. Here it is: https://emmastudies.com/ Sara Pinto Jan 10, 2022, 4:18 PM Mónica Freitas, they have free planners of all kinds and shapes, free calendars, and even good tips on how to be productive! Mónica Freitas Jan 10, 2022, 5:55 PM Sara Pinto, I like that they have a printable section. It makes it easier to navigate. Have you ever used any of their templates? Sara Pinto Jan 11, 2022, 10:23 AM Mónica Freitas, so far, only the calendars and I read all their posts. But I haven't had the chance to use one of the planners yet. Mónica Freitas Jan 11, 2022, 10:43 AM Sara Pinto, do you find the blog layout confusing? I was checking it, but it was hard to find the "really important" posts that aren't just answering other people's comments. Sara Pinto Jan 17, 2022, 11:47 AM Mónica Freitas, I usually don't have an issue since I follow it on other platforms, and the creator shares the direct link. But if you click on "my posts" it might be easier. 🔥 1 Mónica Freitas Jan 17, 2022, 12:10 PM Sara Pinto, thanks for the tip!

Informative Blogs to Follow in 2022

This is the first of the 'How I built' trilogy detailing the build process of all 3 of my clusters. As you may have already read in my , I built my first cluster, in April 2017 for a really superficial reason, if you haven't read it yet, I suggest you do that for a good laugh. earlier post Octopi Origins I chose Octopi, the of the word as it's the only word I could think of with a in it. Coincidentally, the prefix is also a quantifier, signifying 8 and so I thought it would be interesting to create a ludicrous 8-node Raspberry Pi cluster. unsanctioned latin plural Octopus Pi 'octo' The standard pluralised form of "octopus" in English is "octopuses". The alternative plural "octopi" is considered grammatically incorrect because it wrongly assumes that octopus is a Latin second declension "-us" noun or adjective when, in either Greek or Latin, it is a third declension noun. ( ) Wikipedia Lack of budget I had planned to have 8 nodes to fit the name but eventually found it a little too extravagant given that I was a poor student back in 2017 and I had absolutely no idea what to do with the cluster other than to admire its beauty on my desk. In the end, it was technically a as I only managed to procure 5 nodes for cheap off , but the name stuck nonetheless. Pentapi Carousell Cluster specifications |         | 1 x Raspberry Pi Model B+ | 4 x Raspberry Pi Model B |
|---------|---------------------------|--------------------------|
| CPU     | 1× ARM1176JZF-S 700 MHz   | 1× ARM1176JZF-S 700 MHz  |
| RAM     | 512MB                     | 512MB                    |
| USB     | 4x USB 2.0                | 2x USB 2.0               |
| Storage | 16GB MicroSD Class 10     | 8GB SD Class 10          | The first cluster wasn't handsome from the beginning. Before it looked the way you're probably already accustomed to by now, it looked like this: Janky beginnings Not too long after, my neat freak tendencies got the better of me and I bought a laser-cut acrylic stacking-rack for the cluster. There's always an excuse for not enough compute but there's no excuse for an unsightly cluster. But there's always budget for looks Now you may be asking, how much did I spend in total to build Octopi? All in all, I spent building this cluster. Here's the cost breakdown for each component: $183.34 Parts List | Item                                             | Qty | Cost (SGD)   | Cost (USD)* |
|--------------------------------------------------|-----|--------------|-------------|
| Raspberry Pi Model B+ (Used)                     | 1   | S$25.00      | $18.04      |
| Raspberry Pi Model B (Used)                      | 4   | S$80.00      | $57.74      |
| Anker PowerPort 5 40W 5-port USB Charger         | 1   | S$32.58      | $23.51      |
| TP-Link TL-SF1008D 8-Port Ethernet Switch        | 1   | S$20.00      | $14.43      |
| Cat 5e 30cm cable                                | 5   | S$2.50       | $1.80       |
| Cat 5e 1m cable                                  | 1   | S$2.00       | $1.44       |
| MicroUSB charging cable                          | 5   | S$15.00      | $10.83      |
| 8GB Class 10 SD Card                             | 4   | S$24.00      | $36.09      |
| 16GB Class 10 MicroSD Card                       | 1   | S$10.00      | $36.09      |
| Amazon cardboard box                             | 1   | Free         | Free        |
| Rubber band                                      | 4   | Free         | Free        |
| **GeauxRobot Raspberry Pi 5-Layer Dog Bone Stack | 1   | S$41.55      | $29.99      |
| |     | | | SGD/USD exchange rate is 0.72172 as of the time of writing
** Optional but handsome **Total** **S$254.05** **$183.34** * And that was the story of how I broke my wallet in university 3 years back. Provisioning pains For a good half a year, the furthest extent I went was flashing the SD cards with . Essentially, the cluster was a $180 on my desk, consuming 15W to produce heat, in an already hot tropical climate, and light, strong enough to disrupt my sleep cycle. Raspbian kinetic sculpture Then one day, I accumulated enough guilt to motivate myself to stop wasting electricity and start putting the cluster to work for me, and it was only then I started setting the cluster up proper. Initially, I tried provisioning the cluster one by one. With Raspbian already set up, the next thing I had to do was to enable password-less, secure communications from my local machine to the cluster and between cluster nodes. At first, that task sounded simple but I quickly realized how painful this process was. To give you some context, here are the steps I had to follow and repeat 5 times to enable password-less authentication from my machine to each node: 1. Install , a text editor 2. Copy my machine's public key into the file 3. Disable password authentication by editing Generate a public/private key pair on the node in Copy the node's public key and host key into a text editor on my machine vim authorized_keys ~/.ssh/config 4. ~/.ssh/id_rsa 5. After completing the above 5 steps, in order to enable password-less authentication between each node, I had to to: login to every node again 1. Paste the public key of into the file 2. Paste the host key of into every other node except itself authorized_keys every other node except itself /etc/ssh/ssh_known_hosts Those last 2 steps were what I'd call, n² complexity, where the number of manual operations grows to the square of the number of nodes. 😱 I ran out of gas by the time I reached the third node. Enter Ansible I searched around on how I could automate the provisioning steps above, and it was then I picked up , an open-source, software provisioning, configuration management, and application deployment tool. Ansible allows one to write Playbooks, which are essentially a list of tasks in intuitive syntax, to be run on a pre-defined list of target machines. Ansible yaml All of the above tasks that took me 2 hours was done in 10 minutes with a single command: $ ansible-playbook octopi.yml This is how I my first Ansible playbook looked like: octopi.yml --- - hosts: octopi vars: ansible_user: pi comment_re_prefix: '^[# \t]*' # SSH Key Configuration security_ssh_keygen_algorithm: 'rsa' security_ssh_keygen_bits: '2048' security_ssh_keygen_password: '' security_ssh_admin_key: ~/id_rsa.pub # SSH Configuration security_sshd_config_path: /etc/ssh/sshd_config security_ssh_port: 22 security_ssh_password_auth: "no" security_ssh_permit_root_login: "no" security_ssh_usedns: "no" security_ssh_permit_empty_passwords: "no" security_ssh_challenge_response_auth: "no" security_ssh_gss_api_auth: "no" security_ssh_x11_forwarding: "no" tasks: - name: Install common packages become: yes apt: name: vim state: present - name: Check if ssh key exists stat: path: '/home/ /.ssh/id_rsa' {{ ansible_user }} register: security_ssh_key_stat - name: Create .ssh directory file: path: '/home/ /.ssh/' {{ ansible_user }} state: directory - name: Generate ssh key command: 'ssh-keygen -t -b -N " " -f /home/ /.ssh/id_ ' {{ security_ssh_keygen_algorithm }} {{ security_ssh_keygen_bits }} {{ security_ssh_keygen_password }} {{ ansible_user }} {{ security_ssh_keygen_algorithm }} when: security_ssh_key_stat.stat.exists == False - name: Slurp public keys from all nodes slurp: src: '/home/ /.ssh/id_ .pub' {{ ansible_user }} {{ security_ssh_keygen_algorithm }} register: security_ssh_key_slurp - name: Copy public keys of all nodes into authorized_keys authorized_key: user: ' ' {{ ansible_user }} state: present key: ' ' {{ hostvars[item].security_ssh_key_slurp.content | b64decode }} with_items: ' ' {{ ansible_play_hosts }} - name: Copy local public key to authorized_keys authorized_key: user: ' ' {{ ansible_user }} state: present key: " " {{ lookup('file', security_ssh_admin_key) }} - name: Slurp host keys from all nodes slurp: src: '/etc/ssh/ssh_host_ _key.pub' {{ security_ssh_host_key_algorithm }} register: security_ssh_host_key_slurp - name: Insert all nodes into global known_hosts become: yes known_hosts: name: ' ' {{ item }} key: ' ' {{ item }} {{ hostvars[item].security_ssh_host_key_slurp.content | b64decode }} path: '/etc/ssh/ssh_known_hosts' hash_host: ' ' {{ security_ssh_hash_known_hosts }} state: present with_items: ' ' {{ ansible_play_hosts }} - name: Secure SSH configuration become: yes lineinfile: path: ' ' {{ security_sshd_config_path }} regexp: ' ' {{ item.regexp }} line: ' ' {{ item.line }} state: present with_items: - regexp: ' Port' {{ comment_re_prefix }} line: 'Port ' {{ security_ssh_port }} - regexp: ' PasswordAuthentication' {{ comment_re_prefix }} line: 'PasswordAuthentication ' {{ security_ssh_password_auth }} - regexp: ' PermitRootLogin' {{ comment_re_prefix }} line: 'PermitRootLogin ' {{ security_ssh_permit_root_login }} - regexp: ' UseDNS' {{ comment_re_prefix }} line: 'UseDNS ' {{ security_ssh_usedns }} - regexp: ' PermitEmptyPasswords' {{ comment_re_prefix }} line: 'PermitEmptyPasswords ' {{ security_ssh_permit_empty_passwords }} - regexp: ' ChallengeResponseAuthentication' {{ comment_re_prefix }} line: 'ChallengeResponseAuthentication ' {{ security_ssh_challenge_response_auth }} - regexp: ' GSSAPIAuthentication' {{ comment_re_prefix }} line: 'GSSAPIAuthentication ' {{ security_ssh_gss_api_auth }} - regexp: ' X11Forwarding' {{ comment_re_prefix }} line: 'X11Forwarding ' {{ security_ssh_x11_forwarding }} You can get this Ansible playbook at my GitHub repository. I have also written other useful tasks such as disabling the HDMI to reduce the power draw (written out of guilt after having wasted 65kWh over 6 months by letting the cluster idle) or disabling swap to extend the SD card's lifespan. cluster-ansible-playbooks My foray into Ansible was inspired by on his Pi Dramble cluster where he runs Drupal on a cluster of 6 Raspberry Pi 2B. I highly recommend checking his site out if you're just getting into Ansible. Jeff Geerling's blog Ansible was my greatest discovery back then and still remains very much relevant today, not just in my newest cluster but also in my career. After the trauma from provisioning the cluster, I took a break of several weeks to recover (and study for my exams) before starting work on hosting a Drupal blog. I did not end up using the blog as the university network was firewall-ed and not accessible from the outside world not to mention I was a lazy student then. Nevertheless, it was definitely a worthwhile project and I'll explain why with my next post in this series. Should you build this cluster? 1. Are on a really tight budget 2. Are keen on getting your hands dirty to learn how to deploy applications from the ground up 3. Have little to no experience with Linux and networks 4. Wish to explore Docker This build is for you if you: 1. Are already familiar with Linux basics 2. Have experience hosting something like a Wish to learn Kubernetes This build is not for you if you: LEMP stack 3. Should you choose to build this cluster, there are some things that you should be aware of before starting. What you should know There were quite a number of limitations when it comes to using Raspberry Pi Model 1B, most notably, the single-core 700MHz processor which may prove to be under-powered for many tasks. Notwithstanding, I shall list out the other points that I wished I was aware of before I built it. The standby power consumption of Raspberry Pi Model 1B is than that of the Model 1B+ (at 3.5W vs 1W) due to a flaw in the design of their power management hardware . Even though 2W may not sound like a huge difference, it stacks up quickly with 4 of these running 24/7. Therefore, it's almost always advisable to get the Model 1B+ and newer models. Elevated power consumption more than 3x higher (Wikipedia) The storage performance on the Raspberry Pi Models 1 through 3B+ is suboptimal. Even if you were to use the best SanDisk Extreme Pro+ SD card intended for videography usage, it would perform the same as when a standard Class 10 SD card was used. Even on a Class 10 SD card, the I/O (read/write) performance is severely bottlenecked by the shared between the Ethernet Port, the SD card slot and all USB ports. Limited storage speed single USB 2.0 bus USB 2.0 is at its core, a 480mbps connection, enabling full speed communication in only one direction at any instant, effectively halving the bi-directional communication bandwidth to 240mbps. Ethernet on the other hand, is 100mbps allowing 100mbps bi-directional data transfer at any point in time. half-duplex full-duplex, To put things into perspective, in a typical large file download operation, 480mbps of USB 2.0 theoretical bandwidth (~400mbps actual due to protocol overheads) is shared between the I/O operations to your SD card, USB flash drive, and receive/transmit operations to the Ethernet cable. | Operation                    | Read bandwidth | Write bandwidth |
|------------------------------|----------------|-----------------|
| Actual available bandwidth   | 200Mbps        | 200Mbps         |
| Ethernet maximum utilization | -100Mbps       | -100Mbps        |
| USB typical utilization      | -25Mbps        | -25Mbps         |
| | | | **Available for SD Card** **75Mbps** **75Mbps** Tabulating the bandwidth utilizations, we see that only 150mbps of actual bandwidth is available to I/O operations, amounting to a maximum of 9.38MB/s simultaneous read and write speeds, which is slightly below the maximum write speed of 10MB/s in the lowest tier Class 10 U1 SD cards. This is quite a hotly debated issue in the community albeit not very visible to those outside. The justification given by the Raspberry Pi foundation was that using a single shared USB bus keeps the manufacturing costs low, which I agree as most use-cases except mine are not demanding on the I/O aspect. My two cents Despite the restrictive hardware of a single 700MHz CPU and 512MB of RAM per node a surprising number of things can be learnt from it. Besides, most of the time, unless you go into the realm of streaming content and downloading large files, you will not hit the I/O bottleneck. Though instead of 5 nodes, I would recommend building a cluster with 3 nodes as the performance gain of two additional nodes is insignificant as compared to the additional investment it requires. If this is your first exposure to Linux, definitely build yourself one! What's next? In my next post in this series, I shall introduce the Kraken cluster which has since replaced the Octopi cluster, so stay with me! Originally published at https://ikarus.sg on June 19, 2020.

Building Cheap Raspberry Pi Cluster from Scratch

“Botting” is the process of using software automation to perform tasks. It can has become popular on social media. Some companies provide this as a service to boost following on social media. This makes use of “bots” much like the ones used in old IRC chat rooms. Bots can also spam and attack networks when coordinated. They are illegal to use on social media platforms, yet they are still being provided as a service by third parties. Celebrities and many public figures use these services to boost their social media image. It makes them appear to have many followers and that in effect can legitimize being a social influencer. From a business perspective that can be monetized through sponsors who advertise products and services. Bots are software apps that can automate tasks Social media sites are now cracking down on this. Instagram uses a neural network, which can detect botting by analyzing word frequency in comments posted by accounts. If you get caught, your account(s) gets a “shadowban” or even deleted. This is the reason some people’s account suddenly get deleted or deactivated. When Instagram detects anomalous activity from the account, it shuts it down. Some companies provide botting service for a fee. The way this works is the company takes your Instagram or Twitter account and with your permission creates a bot that performs tasks on your behalf. So this means a bot is doing the likes, comments and following as your account, and it can do thousands of these a day. The result is to get more social engagement like followers, comments, likes and the occasional DM. Many who have tried these services see it otherwise. A photographer who tried botting for 2 years had an awkward experience when the bots made unnecessary comments or liked photos which were not even of great quality. Another purpose of bots is to follow other users. Some companies that promoted botting in the past have recently been shutdown. This was upon the request of non other than Instagram. Some sites have shutdown due to botting activities which social media apps don’t allow. Some social media promotion companies actually use your account to follow other accounts and likewise. The tradeoff here is that you get followers, but you also need to follow other accounts. It is all automated as a task. Some followers are obviously fake with names like “GET_More_Followers” or “InstaFollowers4Free”. These are part of a botnet or bot network that are triggered to follow a certain user account. Bots also can reference hashtags like #picoftheday #summertime #YoLo. In Instagram, code automation will then use your account to go over a slide of photos with a certain hashtag and either like them or leave a comment. Sometimes it is easy to tell that a comment is left by a bot when it just sounds strange or inappropriate. Webmasters have implemented a “robots.txt” file on servers to provide a guideline for bots to follow when visiting their website. Any bot that is “spidering” or violating these rules is removed or banned. Twitter and other social media sites have likewise implemented their own “robots.txt” file to prevent bots spamming and fake users, but there are automated tasks that are still able to pass. For bots less risky tasks such as liking, following and unfollowing seem to get through, as happens on IG and Twitter. These bots may not actually be increasing any organic social engagements at all, since they tend to interact with other bots. Another way tasks are automated is with macroing, which uses scripts to perform the things that allow delegation of a user’s account to like, comment, follow and unfollow an acccount. Code can be written to run an automated task for a social media account as shown in this pseudocode. As Instagram states in their “General Conditions Instagram Terms of Use”: 15. You must not create accounts with the Service through unauthorized means, including but not limited to, by using an automated device, script, bot, spider, crawler or scraper. Disclaimer: This information is provided for reference. No actual code automation or botting was used for this presentation. References: Photographer who botted for 2 years: https://petapixel.com/2017/04/06/spent-two-years-botting-instagram-heres-learned/ Instagram Terms of Use: https://www.instagram.com/about/legal/terms/before-january-19-2013/ Botting defined on Wikipedia: https://en.wikipedia.org/wiki/Internet_bot Instagress shutdown: https://www.theverge.com/2017/4/20/15374080/instagram-bot-site-instagress-dead

“Botting” and Task Automation on Social Media

There’s no golden formula when it comes to calculating the perfect price for your new . A lot of times I have seen people pricing their products based on what they think is right, rather than finding out how their users value that product. product When I was in school, one of my friends introduced me to a tycoon called and I spent hours playing it, everyday. I enjoyed it so much that I started looking for more tycoon games. Tycoon games were fun and they taught me how the real world works. How a business works. And how it’s not always a sweet ride. game Lemonade Tycoon Soon I came across , a game by Atari (Edit: It’s actually Chris Sawyer who made the game, which was published by Hasbro Interactive and then Atari licensed it to release more RCT games). Most of you must’ve already heard about it or played it. If not, it’s a game where you’re given a piece of land and your goal is to convert it into an amusement park earning as much profit as possible. Roller Coaster Tycoon I wouldn’t dive much into the details of the game, but if you haven’t played it, then you need to know that a park can offer two kinds of pricing: : The entry to the park is free but guests need to pay for each ride. Pay per ride : You charge an entry fee but the guests can go to any number of rides for free. Pay per entry I preferred to play with the former model. It was more fun to experiment with the pricing of individual rides and observe the guests’ reactions. After spending hours on this game, I observed that no matter how experienced I got, I just couldn’t guess the right price for a ride. I always ended up increasing or reducing the price to allow me to have a balanced profit + guests influx. What does this have to do with the real world? A couple months back, I made an app, “ ”, which lets you pick colors from around you using your iPhone’s camera. Cone, being my first native iOS app, took me about two month to finish. The initial version had a good balance of usability, functionality and delight (some people ignore delight in their MVPs but I believe it’s as important as the other two). Cone Now, I hadn’t made any paid products in the past so I didn’t have any idea on how to price Cone. There were a number of ways—make it free with ads, make it free with in-app purchases, or make it paid with a flat price. I don’t really like when apps show me ads pasted all over their interface, so this option was instantly crossed out. I briefly considered the second option, add in-app purchases, but having recently finished with the development I didn’t want to learn yet another complex thing so I ignored this as well. Finally I decided to price it at without much thought and hit submit. Before submitting it to the App Store, I hadn’t asked anyone about how much they’d pay for it. As a result, I didn’t have any idea if I’d priced it too low or if people would actually pay for it. $0.99 Lessons I learnt and the impact of RCT on pricing Cone got hunted on a day after I had launched it. It was purchased about 200 times and I couldn’t contain my excitement. I was so happy to see that people were actually finding it useful and paying for it. Product Hunt Then, a couple weeks later, I decided to submit it to and the response I got was overwhelming. Again, I received great feedback and about 500 sales from it. But this time, I noticed a pattern which reminded me of Roller Coaster Tycoon. People found Cone extremely useful and I saw comments like “ ”, “ ”, “ ”. Hacker News This is really good value! It’s really cheap for what it does I would’ve easily paid 3x–5x of what you’re charging There are 3 guests’ reactions you normally see in RCT for a ride w.r.t. pricing. 1. [Ride] is really good value Reactions when I set the entry of the ride to $2.00 This is where you know that the price you’ve set is too low. You can increase the price and the guests will still keep paying for it. 2. [Ride] was great Reactions when I set the entry of the ride to $8.00 This is a sweet spot (value = money). You can still experiment with the fee and increase it a bit to see if any guests are running away or not. 3. I’m not paying that much to go on [ride] Reactions when I set the entry of the ride to $15.00 This means you’ve overpriced the ride and the guests don’t see value for money. You start reducing the price here and try to reach the sweet spot. I started seeing satisfactory results at $12. Now, you see, it works the same in the real world. Well, in most of the cases. Once I started hearing from people that Cone is really good value, and that people are ready to pay more, I started experimenting with the pricing. I silently increased the price to and kept promoting it like normally. To my surprise, people still believed it’s good value (!!!). Here’s the graph showing number of downloads and the price of the app. $1.99 That PetaPixel spike was probably the happiest day of my life As you can see, the price change didn’t really affect the downloads. In turn, people were still finding it great value for money. Which means that I was still at stage 1 of reactions according to RCT, and there’s still some scope of increasing the price further to get to the sweet spot. A couple days back (27th July, 2017 to be exact), I increased the price by a dollar, making it . And guess what? People still found it good value for money. For now, I am going to sit at this pricing and observe more reactions before experimenting further. There’s a long way to go before reaching that second stage. $2.99 Conclusion Roller Coaster Tycoon taught me that it’s difficult to put a correct price tag on a product. You have to keep experimenting to get the price right. Start with a low price and work your way up to the point you stop acquiring new customers. Then take one step back. That’s the correct price tag for your product. I’ll keep updating my progress with the experiment on . Twitter

Stories, News & Mentions about PENTOPIX

Pentopix’s Stories on HackerNoon