pookieboo

In deep learning, dozens of hyperparameters influence how effectively a model learns — but none is more critical than the learning rate. Whether training proceeds smoothly, crawls painfully, or explodes into chaos depends largely on this single value. A model can have the perfect architecture, the right optimizer, and a clean dataset, but with a poor learning rate, everything falls apart.

In this blog, I explore the learning rate from both a theoretical and experimental lens. After explaining what it is and why it matters, I run a simple experiment using a small CNN on the MNIST digit dataset and train it with three different learning rates. By comparing how the same model behaves under different settings, we see how dramatically this hyperparameter shapes the entire training process.

The aim is not just to explain what the learning rate is, but to understand how it shapes the learning process — slow, fast, or downright furious.

1. What Exactly Is the Learning Rate?

The learning rate, commonly represented as η, is a hyperparameter that controls how much a neural network’s weights are updated during training. Each update step aims to reduce the model’s loss, and the learning rate determines how large or small that step should be.

At the core of deep learning optimization is gradient descent. During each iteration, the model computes the gradient of the loss function with respect to its parameters and adjusts those parameters in the opposite direction of the gradient.

Mathematically, weight updates follow the rule:

This equation highlights the purpose of the learning rate.

Purpose of the Learning Rate

Controls update size: A high learning rate results in large weight updates, while a low one leads to small updates.

Determines convergence behavior: It influences how quickly the model learns, how stable training is, and whether the model reaches an optimal solution.

Balances speed and stability: The learning rate must be high enough to make meaningful progress but low enough to avoid overshooting the optimum.

Theoretical Effects on Training

1. Effect on Model Convergence

The learning rate directly influences how fast and how smoothly the model converges to a minimum of the loss function.

Too high: The model may overshoot minima, causing oscillation or divergence.

Too low: Training becomes extremely slow and may get stuck in flat regions of the loss surface.

Well-tuned: The model converges efficiently with stable decreases in loss.

2. Effect on Model Performance

A poorly chosen learning rate impacts performance significantly:

High LR → unstable gradients → poor accuracy

Low LR → underfitting due to insufficient progress

Optimal LR → best balance of exploration + stability

3. Effect on Generalization

Generalization refers to how well the model performs on unseen data.

If LR is too high, the model may never settle into a good minimum and ends up with high test error.

If LR is too low, the model may converge to a suboptimal minimum or overfit due to slow learning.

Proper LR helps the optimizer find a region of the loss landscape that generalizes well.

Experiment: Testing the Effect of Learning Rates on CNN Training

To understand how the learning rate influences the behavior of a deep learning model, I conducted a controlled experiment using a simplified convolutional neural network trained on a reduced subset of the MNIST dataset. By keeping the model architecture, optimizer, and number of epochs constant—and changing only the learning rate—we can clearly isolate its impact on convergence and performance.

Objective:

The goal of this experiment is to observe how three different learning rates affect:

• training speed • stability of the loss curve • accuracy progression • overall learning behavior

This helps demonstrate, in a concrete and visual way, why the learning rate is considered one of the most important hyperparameters in deep learning.

Dataset

To make the experiment lightweight and fast while still meaningful, I used a subset of the MNIST dataset:

10,000 training images

2,000 test images

Grayscale, 28×28 resolution

10 class labels (digits 0–9)

MNIST is widely used for introductory experiments because it is clean, simple, and efficient to train on—even with modest hardware.

Model Architecture

For this experiment, I used a compact CNN designed specifically for speed while retaining enough capacity to learn digit features. The model consists of:

Conv2D (1 → 8 filters), kernel size 3

ReLU activation

Max Pooling (2×2)

Conv2D (8 → 16 filters), kernel size 3

ReLU activation

Max Pooling (2×2)

Flatten layer

Fully Connected Layer (400 → 64 units)

ReLU activation

Output Layer (64 → 10 units)

This smaller CNN trains in seconds, making it ideal for quick comparisons across multiple learning rate settings.

Learning Rates Tested

I trained the exact same model with three different learning rates:

0.1

0.01

0.0001

5. Results

After training the same CNN with three different learning rates (0.1, 0.01, and 0.0001) for 3 epochs on a 10,000-sample MNIST training subset and evaluating on 2,000 test images, I recorded the loss and accuracy per epoch.

Numerical summary

LR = 0.1

Loss: 3.1660 → 2.3058 → 2.3066

Accuracy: 11.7% → 8.9% → 11.7%

LR = 0.01

Loss: 0.5158 → 0.1239 → 0.0872

Accuracy: 94.25% → 96.20% → 96.20%

LR = 0.0001

Loss: 2.2539 → 2.0417 → 1.6081

Accuracy: 43.20% → 60.50% → 66.80%

Figure 1 (Training Loss vs Epochs) shows that:

The model with LR = 0.01 rapidly drives the loss down close to zero within just three epochs.

The LR = 0.0001 model also reduces loss, but much more slowly and remains relatively high after three epochs.

The LR = 0.1 model’s loss decreases initially but then plateaus around ~2.3, indicating that it fails to continue learning effectively.

Figure 2 (Test Accuracy vs Epochs) shows that:

LR = 0.01 reaches around 96% test accuracy, stabilizing by the second epoch.

LR = 0.0001 improves steadily from about 43% to 66.8%, but remains far behind LR = 0.01 within the same number of epochs.

LR = 0.1 hovers around 10–12% accuracy, which is close to random guessing for 10 classes, meaning the model essentially fails to learn.

6. Analysis: Slow, Fast, or Furious?

These results line up very well with the theoretical expectations discussed earlier.

LR = 0.1 — Furious and Unstable

The highest learning rate behaves like the “furious” setting:

Accuracy stays near chance level (around 10–12%), even after three epochs.

Loss briefly drops from 3.16 to 2.30 but then stops improving, suggesting that the optimizer is taking steps that are too large to move toward a good minimum.

This matches the theory that an excessively high learning rate can cause the optimizer to overshoot minima and effectively bounce around without settling.

In practice, this learning rate is clearly too aggressive for this model and dataset.

LR = 0.01 — Fast and Effective

The middle learning rate, 0.01, clearly wins in this experiment:

Training loss drops sharply from 0.51 to 0.08 in just three epochs.

Test accuracy quickly rises above 94% in the first epoch and stabilizes at around 96.2% by the second epoch.

The loss and accuracy curves are smooth and stable, indicating healthy convergence.

This setting represents the “fast but controlled” regime: the model learns quickly without becoming unstable. It also shows that a good learning rate can reach strong performance in just a few epochs, saving both time and compute.

LR = 0.0001 — Slow but Safe

The smallest learning rate behaves like the “slow” setting:

Loss decreases gradually but remains relatively high after three epochs.

Accuracy improves from 43.2% to 66.8%, which is clearly better than random but still much lower than the 96.2% achieved by LR = 0.01.

The curves are stable and monotonic, indicating that the model is learning, just very slowly.

This reflects the theoretical behavior of a too-small learning rate: safe, stable progress but inefficient learning that may require many more epochs to reach competitive performance.

Overall insight

With everything else held constant (same CNN, optimizer, dataset subset, and epochs), the learning rate alone produced three very different behaviors:

Too high (0.1) → almost no learning, unstable training.

Too low (0.0001) → clear learning, but slow and incomplete within the same time budget.

Well-tuned (0.01) → fast, stable convergence and the best accuracy.

This experiment illustrates why the learning rate is often called the most important hyperparameter in deep learning: a good choice can turn the same model from “random guessing” into “highly accurate” with no other changes.

7. Conclusion

In this mini-project, I explored the effect of the learning rate on a small CNN trained on a subset of the MNIST dataset. By keeping the model, optimizer, and data fixed and varying only the learning rate, I observed three distinct regimes:

A furious learning rate (0.1) that was too aggressive and prevented the model from learning.

A slow learning rate (0.0001) that allowed gradual learning but failed to reach high accuracy within a small number of epochs.

A fast but stable learning rate (0.01) that achieved around 96% test accuracy in just a few epochs.

These findings reinforce the theoretical view that the learning rate directly shapes convergence speed, training stability, and final performance. In practical terms, this suggests:

Never rely on a single default learning rate without checking its behavior.

Start with a moderate value (e.g., 0.001–0.01 for Adam) and adjust based on loss/accuracy curves.

If loss barely changes, the learning rate is probably too low; if loss is unstable or accuracy stays near random, it may be too high.

Overall, this experiment made the abstract concept of “learning rate tuning” much more concrete: a single scalar hyperparameter can decide whether a neural network learns slowly, learns efficiently, or doesn’t learn at all.

References

TensorFlow. Learning Rate Schedule Documentation. https://www.tensorflow.org/api_docs/python/tf/keras/optimizers/schedules/LearningRateSchedule

PyTorch. Optimizers – Learning Rate and Parameters. https://pytorch.org/docs/stable/optim.html

Keras. Optimizers in Keras. https://keras.io/api/optimizers/

Brownlee, Jason. Learning Rate for Deep Learning Neural Networks. Machine Learning Mastery. https://machinelearningmastery.com/learning-rate-for-deep-learning-neural-networks/

Towards Data Science. Understanding Learning Rates and Their Effect on Training. https://towardsdatascience.com/understanding-learning-rates-and-how-it-improves-training-5c7b054ca35b

Customer Insights from Continued MVP Testing

In Week 11, we conducted another round of MVP testing to gather more feedback from users—including students, teachers, and professionals—after letting them try out the CareXpress platform. This week, the focus was on improving usability, prescription tracking, map and location accuracy, and feature expectations.

Click here to view Week 11 Documentation.

Key Takeaways from User Interviews

Ease of Use

Most users found the app easy to navigate due to its straightforward layout.

Some users from older generations still found it hard to navigate and suggested tutorials for first-time users.

Prescription Handling and Security

Users appreciated that digital prescriptions couldn’t be screenshotted or altered, increasing trust.

There was a suggestion to retain expired prescriptions in a read-only format for future medical consultations.

Pharmacy Listings and Medicine Descriptions

Many noted that pharmacy listings were limited and lacked complete inventory data.

Users emphasized the need for medicine descriptions, including dosage, purpose, and use-case scenarios.

Map Accuracy and Tracking Issues

Multiple users reported that the map did not show real-time locations and sometimes displayed incorrect user positions.

Users wanted rider tracking and communication features, not just static distance displays.

Login and Profile System

Users found the login page frustrating. They suggested enabling phone number login, like Grab or Foodpanda.

There was also a request for role-specific options (customer-only vs. patient-user).

General Bug Reports and UI Issues

Browse button was unresponsive for some users.

Font contrast issues made some text hard to read.

Users suggested adding pharmacy branch names to help with selection.

Decisions Made This Week

To address these concerns, the following updates will be prioritized:

Improving the Map System – Fixing the real-time tracking issues and correcting location inaccuracies.

Prescription History Access – Making expired prescriptions viewable in a read-only mode for medical referencing.

Adding Medicine Descriptions – Providing users with detailed information per product.

Enhancing Login Options – Implementing phone number authentication and allowing role selection.

Rider Tracking and Communication – Enabling live tracking of riders and a messaging system between rider and customer.

UI/UX Enhancements – Fixing browse button, improving text readability, and showing branch names.

Lessons Learned

Week 11 reaffirmed that clarity, security, and accessibility are the pillars of trust in health-tech apps. Users appreciate simplicity, but also want robust features that meet everyday needs—like detailed medication info, accurate tracking, and easy access. Addressing technical gaps and supporting both tech-savvy and non-tech users will be essential as we prepare the next version of CareXpress.

Hanggang sa Muli,

Customer Insights and MVP Evaluation

This week marked a milestone for CareXpress as we released our Minimum Viable Product (MVP) and conducted interviews with users who interacted with the platform. The goal was to test usability, security, clarity of features, and gather suggestions for improvements.

Click this link to visit our Week 10 Documentation.

Key Takeaways from MVP User Interviews

Ease of Use – Most users found the app intuitive and easy to navigate. Students and riders especially appreciated the convenience of digital prescriptions. However, there were concerns about accessibility for elderly users.

Prescription Security & Convenience – Users felt secure knowing prescriptions were stored in-app and couldn’t be altered or easily shared. They noted it's more readable and reliable than paper prescriptions.

Performance Concerns – Some users experienced slow loading times, especially when internet connectivity was poor. Offline support or faster performance was suggested.

Feature Suggestions

Refill reminders and order tracking were highly appreciated.

Users requested medicine descriptions to better understand what they’re ordering.

Display of generic vs branded options and pharmacy stock visibility was requested.

A customer service contact or consultation chat was seen as a helpful addition.

Pharmacy and Rider Insights

Users preferred seeing more pharmacy listings, especially large chains.

Riders recommended improvements to the driver interface such as route previews, order distance, profit estimate, and payment breakdown.

A "pick order" feature for riders was suggested.

Decisions Made This Week

Based on the MVP feedback, we are initiating several changes to improve the user experience and build trust.

Key Adjustments

Optimizing App Speed – Improving loading times and exploring offline caching for prescriptions.

Adding Medication Descriptions – Including clear info on medicine names, uses, and differences between branded/generic.

Enhancing Pharmacy Listings – Expanding the directory to include more big-name pharmacies.

Refining Rider Interface – Allowing riders to view route, order distance, and profit before accepting.

Expanding Support – Adding help/chat feature and clearer instructions for QR codes and app navigation.

Lessons Learned

The MVP testing confirmed that users value convenience, security, and clarity. While the app’s foundation is strong, there's room for improvement, especially in areas like elderly usability, order tracking, and rider experience. Continuous feedback and iteration will be critical as we move toward the next version of CareXpress.

Bye-bye,

Expanding Our Understanding Through Customer & Pharmacist Interviews

In Week 9, we broadened our research by interviewing both medicine buyers and pharmacy professionals. Our goal was to understand the real-world obstacles on both sides—how customers manage prescriptions and how pharmacies handle orders, especially in the absence of dedicated digital platforms.

To view our Week 9 Documentation, kindly visit this link.

Key Takeaways from Interviews

Long Wait Times and Stock Issues – Customers consistently mentioned long lines and delayed service, especially when pharmacies needed to check availability across branches. Stock unavailability remains a common frustration.

Prescription Delivery Gaps – Most customers are not yet using medicine delivery services, mainly due to a lack of awareness or availability in their area. However, there's strong interest, especially from solo dwellers and busy workers.

Pharmacists’ Frustration with Fake or Illegible Prescriptions – Pharmacists frequently encounter outdated or hard-to-read prescriptions, especially handwritten ones. This causes issues in processing and leads to customer dissatisfaction.

Limited Use of Digital Systems – Some pharmacies only use phones or messaging apps for orders. Others, like Southstar, use software for inventory but still rely on manual logs for prescription tracking.

Open But Cautious About Tech Adoption – Pharmacists are generally open to using delivery platforms but remain concerned about monitoring prescription validity, especially with digital prescriptions.

Key Adjustments Based on Feedback

Scheduled Delivery Feature – We’re planning a flexible delivery option that allows users to book a timeslot before or after work, especially helpful for solo dwellers and professionals.

Prescription Validation Layer – CareXpress will incorporate prescription expiry checks and flag reused prescriptions, ensuring accuracy for pharmacies and safety for users.

Reliable Pricing & Comparison – While pharmacists are cautiously open to price visibility, we’re exploring ethical and user-friendly ways to compare options without overwhelming users.

Lessons Learned

This week reminded us how much both pharmacists and customers value clarity, speed, and trust. By solving real problems like fake prescriptions, long waits, and lack of coordination, CareXpress can truly stand out. Listening to both ends of the medicine journey—buyers and providers—is key to making this work.

Shalom,

Pharmacy Feedback & System Observations

This week, our focus shifted to pharmacy-side insights to better understand the operations, challenges, and openness to third-party delivery platforms. We interviewed pharmacy assistants, staff riders, and individuals managing in-house deliveries across several locations.

For our Week 9 Interview Documentation. Kindly visit this link.

Key Takeaways from Pharmacy Interviews

Third-Party Platform Reliance – Many pharmacies rely on Grab or Foodpanda for medicine delivery. While convenient, they face issues like outdated inventory syncing and lack of control over delivery timing.

Prescription Verification Practices – Pharmacists remain strict with prescription validation. Most accept digital prescriptions via Viber or Messenger but enforce expiry checks, especially for antibiotics.

In-House Delivery Constraints – Smaller pharmacies often rely on internal staff or a single rider for deliveries, creating strain during busy hours or in remote areas. Delays and communication gaps with third-party riders (e.g., Maxim) were commonly cited.

Low to No Delivery Charges – Most pharmacies maintain standard medicine prices and offer free delivery for bulk orders. Delivery charges, when applied, range from ₱60–₱100 depending on location.

Digital Communication Gaps – Orders are primarily placed via calls or messages. Most prescriptions are sent and stored in group chats, which presents data management and compliance concerns.

User Perspective Highlight: Living Alone and Needing Faster Access

A solo-living freelance worker shared the recurring struggle of accessing medicine when sick. He recounted dragging himself to the pharmacy during a migraine and missing doses due to lack of delivery options.

Key Adjustments Based on Feedback

Pharmacy Inventory Integration – To address out-of-stock frustrations, we’re exploring ways to sync pharmacy inventory with the app in real time.

Prescription Storage Compliance – We plan to create a secure, centralized prescription system to ensure safe storage and access, replacing informal chat-based records.

Flexible Delivery Fees & Subscription Options – We’re modelling delivery charges in the ₱50–₱80 range and exploring bundled plans for regular users.

Emergency Delivery Protocol – A critical feature in development is the ability to flag orders as urgent, prioritising both dispatch and pharmacy response time.

Lessons Learned

This week showed us how important it is to make things easier for both users and pharmacies. People want medicine delivered fast, safely, and without hassle—especially when they’re sick. Pharmacies, on the other hand, need help managing deliveries and prescriptions properly. By listening to both sides, we can build CareXpress to truly solve real problems and support the people who need it most.

Khodâhâfez,

Customer Insights and Assumptions Tested

As we continue refining CareXpress, this week’s interviews provided further validation and new perspectives on medication accessibility, financial constraints, and potential trust issues with medicine delivery services. We gathered insights from individuals living alone, working professionals, and those caring for family members with maintenance medication needs.

Click here for Week 7 Documentation

Key Takeaways from Customer Interviews

Trust & Pharmacy Preference – Some interviewees prefer buying medicine in person rather than relying on delivery, due to concerns about handling and accuracy.

Challenges for Caregivers – Those buying medicine for family members often struggle with long queues, memorizing prescriptions, and making brand substitutions without proper knowledge.

Access & Affordability Issues – While some respondents live close to pharmacies, others face difficulties due to financial constraints or lack of nearby drugstores.

Delivery Service Potential – Many respondents see value in a medicine delivery service, especially for emergency situations or when they are too sick to go out.

Payment & Digital Adoption – Cash remains the most common payment method, but some users are open to digital payments if it ensures faster transactions.

Key Adjustments Based on Feedback

Building Customer Trust – We are exploring ways to increase trust in our service by ensuring proper pharmacy partnerships, secure transactions, and prescription verification.

Enhancing Caregiver Support – Since many users buy medicine for family members, we are considering features like prescription history tracking and refill reminders.

Flexible Delivery & Payment Models – To accommodate different financial situations, we are refining our pricing structure and allowing both cash-on-delivery and digital payment options.

Marketing for Emergency Use – While some users hesitate to use delivery services regularly, positioning CareXpress as a reliable option for urgent needs can encourage adoption.

Lessons Learned

This week’s interviews reinforced the need to address trust, affordability, and convenience in our service design. While many users are open to digital solutions, some still prefer traditional methods due to habit or security concerns. By emphasizing reliability and accessibility, we can bridge this gap and make CareXpress a viable option for more users.

Nos vemos,

Customer Insights and Assumptions Tested

This week, we continued gathering customer insights to better understand how people buy and manage their medications. Our interviews focused on students, parents with maintenance medication needs, and individuals living alone. These conversations provided deeper insights into the challenges and preferences of our potential users.

Click here for Week 6 Documentation

Key Takeaways from Customer Interviews Medication Accessibility Issues – Many respondents struggle with pharmacy availability, long queues, and finding specific medicine brands. Challenges for Those Living Alone – Individuals who live alone often find it difficult to buy medicine when sick, leading to delays in medication intake. Financial & Time Constraints – Some customers, especially students and working individuals, delay medication purchases due to busy schedules and budget limitations. Growing Interest in Delivery Services – Many interviewees expressed interest in a reliable medicine delivery service, comparing it to existing food and grocery delivery apps. Tracking & Payment Preferences – Most users manually track their medications and prefer a mix of cash, card, or digital payment options like GCash for transactions. Solo Dwellers' Struggles – One interviewee highlighted that living alone makes it difficult to stock up on medicine. They often forget to buy in advance, and when they do get sick, they struggle with getting their medication on time. Delivery services could help ease this burden.

Decisions Made in the Previous Week

Our team analyzed these insights and made further refinements to our CareXpress concept, ensuring it aligns with real user needs.

Key Adjustments Based on Feedback:

Enhancing Convenience Features – Since accessibility is a major concern, we are considering features like pharmacy locator tools and pre-order options to improve availability.

Targeting Solo Dwellers & Busy Professionals – Given their struggles in purchasing medicine when sick, our marketing strategy will highlight the convenience of doorstep delivery for these users.

Flexible Payment & Subscription Models – To cater to different financial situations, we are exploring various pricing models, including per-order fees and possible subscription plans.

Building Trust Through Verification – As prescription accuracy remains a concern, we are looking into verification mechanisms to ensure safe and secure transactions.

Lessons Learned

This week reinforced the importance of understanding user pain points and designing a service that truly meets their needs. While many customers are still used to buying medications in person, the growing demand for convenience presents a great opportunity for CareXpress. Moving forward, we will refine our implementation strategy based on these insights.

toodle-oo,

Week 5 was a crucial milestone for our team as we presented our pitch deck to a panelist during midterms. We received valuable feedback on our presentation, highlighting areas for improvement. Alongside this, we continued interviewing potential users to refine our understanding of their pain points and expectations from CareXpress.

Key Takeaways from Customer Interviews

Challenges in Medication Access – Many respondents experience long wait times in pharmacies, with some waiting for their pension before making a purchase.

Dependency on Physical Visits – Several customers prefer visiting pharmacies in person, either as part of their routine or due to concerns about prescription accuracy in online orders.

Financial Barriers – A recurring issue was the affordability of medicines, leading some to delay purchases or opt for cheaper alternatives.

Openness to Delivery Services – While some respondents were hesitant, many expressed interest in using a delivery service, especially when sick or unable to visit a pharmacy.

Digital Payment & Tracking – Cash remains the primary mode of payment, but a few customers see value in digital payment options and automated refill reminders.

Click here to view our Week 5 Documentation.

During our pitch, the panelists provided essential feedback, prompting us to refine our approach. The key takeaways from their feedback included: Revising Our Revenue Model Slide – Our initial revenue slide was designed as a circular diagram, implying equal revenue sources, which was incorrect. We are now restructuring it to reflect realistic, varying revenue streams. Correcting the Traction Slide – The panel pointed out inconsistencies in our traction slide. We are now revising it to better showcase our progress and key milestones. Enhancing Visuals & Clarity – Similar to previous feedback, we are refining our slides for better readability and engagement.

Lessons Learned

This week reinforced the importance of continuously refining our business pitch. Receiving expert feedback helped us recognize flaws we hadn’t noticed and gave us a clearer direction for improvement.

Our customer interviews also validated our assumptions on financial constraints, accessibility issues, and the growing need for convenient healthcare solutions.

As we move forward, we will continue refining our pitch and adapting our business model to better align with market needs.

La revedere,

In Week 4, we presented our first version of the CareXpress pitch deck. This stage was crucial as we aimed to refine our business proposal based on the insights we received. Through customer interviews and discussions, we tested assumptions about our branding, messaging, and overall presentation effectiveness.

Key Takeaways from the presentation Feedback

Brand Identity & Logo – One of the most critical pieces of feedback was that our initial logo closely resembled an existing company’s logo. This prompted us to redesign our branding to ensure uniqueness and avoid potential legal issues. Clarity & Conciseness – Our pitching script contained some redundant points that made the presentation longer than necessary. We received recommendations to streamline our script for a more impactful pitch. Visual Improvements – The feedback indicated that our slides could be more visually engaging and structured to better convey our business model and customer pain points.

Our new redesigned logo.

New Insights from Week 4 Interviews

Many customers experience delays in refilling medications due to stock shortages and financial constraints. Some wait for their pension or employer-provided medicines.

Pharmacy accessibility remains a challenge, especially in remote areas where only generic medicines are available.

Seniors are open to delivery services but worry about prescription accuracy and package verification.

Manual tracking of medication is common, but digital reminders could help improve consistency.

A pharmacy price comparison feature is appealing but may not be widely used if switching stores is inconvenient.

Click here to view our Week 4 Documentation.

Lessons Learned

This week reinforced the importance of continuous iteration. Customer feedback played a crucial role in helping us refine our pitch and positioning. We learned that clarity, originality, and engagement are key to making a strong impression. Moving forward, we aim to keep iterating based on feedback and ensure that our message resonates effectively with our audience.

Sayonara,

Customer Insights and Assumptions Tested

As we progressed into Week 3 of our customer discovery journey, we continued our interviews, gathering valuable insights into the pain points and needs of our target audience. Through these conversations, we tested several assumptions about the demand for CareXpress, our delivery service for medicines and healthcare products.

Click here for our Week 3 Documentation.

Key Takeaways from Customer Interviews

Convenience and Accessibility – Many customers emphasized the need for a reliable medicine delivery service, especially among elderly individuals and those with chronic illnesses who struggle with mobility.

Trust in Service Providers – Potential users were concerned about the authenticity of medicines and whether the platform could partner with trusted pharmacies.

Delivery Speed and Reliability – Customers wanted assurances that urgent medical needs could be met with fast and efficient delivery.

User-Friendly Interface – Some interviewees expressed concerns about the ease of use, particularly for elderly users who may not be tech-savvy.

Pricing Concerns – While there was interest in the service, cost-effectiveness was a crucial factor in their willingness to adopt the platform.

Decisions Made in the Previous Week

One major decision we made as a team this week was how to structure our first pitch deck and pitching script. With our customer insights in mind, we needed to refine our messaging to ensure that our value proposition was clear and compelling.

First draft of our logo and pitch deck PowerPoint presentation.

Final Choice & Rationale

We decided to focus on a combination of convenience and trust. Our pitch deck would lead with the problem statement, illustrating the challenges faced by our target users, followed by our solution and how we differentiate from existing delivery services. By including testimonials from our interviews, we aimed to make the pitch more relatable and data-driven.

Lessons Learned

Customer Validation Matters – Our initial assumptions were partially correct, but real customer insights helped us refine our approach.

Clear Communication is Key – While preparing the pitch, we realized the importance of being concise yet informative in conveying our business value.

Iterate Based on Feedback – After testing our script, we saw areas for improvement and revised it accordingly.

As we move forward, we plan to refine our pitch further based on feedback from our mentor.

Auf Wiedersehen,

The second round of interviews has begun, and we are uncovering even deeper insights from our targeted participants. As we speak to more people, we are identifying the most common challenges they face when accessing medications while also filtering out concerns that may not be as significant as we initially thought.

Buying medicine should be easy, right? You walk into a pharmacy, pick up what you need, and head home. But in reality, it’s not always that simple. Whether it’s long lines, out-of-stock medications, or just feeling too unwell to make the trip, getting medicine can be more of a hassle than it should be.

Talking to people about their experiences buying medicine, I realized it’s a bigger struggle than it should be. For example, Nethan, one of our interviewees stated that the last thing he wanted to do when sick is leave his house (I mean nagsakit na gani ka ikaw pa magpalit ug medicine diba?)

For another interviewee, Ondo, who has mobility issues, it’s even tougher. Pharmacies may have priority lanes, but they still get crowded.

Not familiar yet with what I'm talking? Here's a picture.

So why isn’t medication delivery more common?

Most people I spoke to said they’d use a delivery service but had concerns. Would they get the right medicine? Would their personal info be safe? How much extra would they have to pay?

Angeline, another interviewee, a pharmacy student, also raised another good point—without face-to-face interaction, people might miss out on pharmacist advice she said.

It is indeed a journey of learning.

All our interviewees enlightened us on a new path on the trajectory of where CareXpress leads. It was like a learning process, as they speak we understand deeply what people truly need. Some assumptions we initially had were challenged, and we realized that some issues we thought were significant might not be as pressing as we believed. It was a good thing nga nag interview jud haha.

Click here to view our Week 2 Interview Documentation.

Some key insights I was able to pickpocket based on the interviews.

Convenience is everything. Mas pilion jud sa tao unsa ang makapasayon sa ila kinabuhi, bakit man gud maglisod lisod sad.

Trust is a big factor. Many people worry about receiving the wrong medicine or sharing their personal health information online. Well since we are dealing with prescription medicines it is important to know whose who are handling the prescriptions.

Affordability matters. While most people are willing to pay a small fee for convenience, there’s a limit to how much they’re willing to spend. Convenience may top our list however the drawback would be if our users are willing to spend another amount for this convenience they want.

This is for the top three on my board list and I wish to uncover more on the next week of our round of interviews.

Konting pahabol, this week wasn’t just about interviews—it was also the week we created our DFV Canvas and Business Thesis. Attached herein is my DFV Canvas (wow murag email kay i-attach).

Conducting Virtual Interviews via Zoom

In our first round of customer discovery, we conducted a series of interviews via Zoom to gain valuable insights into consumer behavior, preferences, and pain points. These interviews focused on various sectors, including agriculture, emergency response, healthcare, transportation, and digital payments. Our goal was to identify potential market gaps for innovative solutions.

Click here to visit our Week 1 Interview Documentation.

Key Findings from Our Customer Discovery

Fresh Produce & Supporting Local Farmers:

Many consumers are open to purchasing cosmetically imperfect produce if it is fresh and affordable. However, trust and quality assurance remain significant factors influencing their purchasing decisions. A direct-to-consumer platform connecting farmers with consumers was well-received, provided it ensures freshness, fair pricing, and convenience.

Emergency Alerts and Response Systems:

Most respondents rely on social media and word-of-mouth for real-time emergency alerts. However, trust in these sources varies, and there is a need for a reliable, government-recognized alert system. Many emphasized the importance of user-friendly interfaces and frequent notifications to ensure responsiveness.

Medicine Accessibility and Delivery Services:

Medicine delivery is seen as a valuable service, particularly for those with health conditions. However, trust issues remain a barrier, with some preferring in-person pharmacy visits despite long queues. Accessibility and stock availability are key factors.

Digital Payments in Public Transport:

While digital payment methods like GCash are commonly used, they are not yet a habit for transport fares. Many commuters experience long queues in bus terminals and are open to a booking system but recognize that it may not fully eliminate congestion. Our Next Steps:

These insights will guide our next phase of research and development. We plan to refine our proposed solutions based on the concerns and preferences raised during these interviews. As we move forward, we will continue conducting more in-depth research to validate and refine our ideas further.

Excerpt of our brainstorming session conducted through Miro.

We all know that in personality assessment tests, there are no right or wrong answers, only shades of in-betweens. When I took the Berkeley Innovation Index test, the results caught me off guard. Although I have never considered myself creative and artistic enough, I was still surprised by how low my innovation index results were. For a moment, I found this question creeping through the back of my mind, "Am I dumb?"

Index Test Results

But Instead of keeping my hopes low, I chose to view these results as an opportunity for growth and a glimmer of hope. There is always room for improving areas in my life that spark personal growth and productivity. Plenty of time to become the best version of myself.

Based on these results, It seems that I have subsided results on Trust, Diversity, Collaboration, and Innovation Zone. While I have positive marks on Resilience, Belief, Allocation, and Comfort Zone. The scores highlight areas of both strength and potential growth. My high scores in Resilience and Belief indicate a strong ability to persevere through challenges and maintain confidence in my ideas. On the other hand, the lower scores in Trust and Collaboration indicate a need to build stronger connections and improve teamwork dynamics. With continued effort (God help me), I am hopeful to see positive growth in these results by the end of the academic term.

Personality Test Results

One thing I've noticed with my results of the 16 Personality Test is that it is ever-changing. Does this mean I undergo phases in my life that alter my personality? Or is it because my frontal lobe is still developing? Well, who knows right?

I still remember the first time I took the test and got an INFJ, then two years later, I retook the same test and became an ISTP, and then ISTP turned into INTJ-T. The one constant, though, is that I am forever Introverted lol. Kidding aside, my personality results continue to evolve, reflecting different stages of my growth and self-awareness.

The results reveal that I lean toward being 53% introverted, which places me right in the middle between introversion and extraversion. My 63% intuitive score and 65% thinking percentage reflect my inclination toward rationality and objectivity. Then the 74% Judging score which greatly resonates with me, as I’ve always been organized and thorough with my life. You can't see me without a to-do list. That is embedded within my system 😆. And lastly the highest of them all, the 89% turbulent trait which highlights my sensitivity to self-improvement.

The journey of self-discovery is indeed ever-changing, and I hope that no matter where the winds of life take me, I am happy and fulfilled.