Portable Modular Power Bank
Abstract
Our project explores the design and feasibility of a modular portable battery pack that allows users to customize and expand storage capacity according to their needs. Unlike traditional fixed-capacity power banks, this design introduces flexibility through attachable modules, enabling users to adapt for everyday use or extended travel without carrying multiple separate chargers. By analyzing market trends, product attributes, and consumer preferences, this study aims to identify whether modularity creates a meaningful advantage in convenience, value, and adoption compared to existing portable chargers. Results show that adding more modules lowers willingness to pay, meaning modularity is not valued by consumers. Fast and wireless charging are the strongest positive drivers, while capacity has little effect and lighter weight is only slightly preferred.
Introduction
The portable modular battery pack is designed to be more compact, flexible, and convenient for everyday use. The magnetic connection system allows the consumer to have the ability to add or remove a battery module, making it easy for one to increase the charging capacity or swap out battery modules that have run out of power. When using the modular battery, the consumer has various charging options for their device, including the built-in USB-C and Lightning ports or through wireless charging, which eliminates the need for the consumer to carry their own cords. Each module also features a digital display showing the remaining power. With the product containing all of these features, the portable modular battery pack offers consumers a customizable and efficient way to keep small devices powered throughout their daily lives.
Survey Design
When creating the pilot survey, several questions were included to determine the eligibility of the respondents. These questions asked whether the participant owned a powerbank, if so the survey would proceed to ask follow-up questions about how often they carry it and if they were currently shopping for a new one. If the respondent does not own a powerbank, the survey would proceed to ask if they own a small electronic device and if they would be interested in purchasing a powerbank in the future. This structure provides flexibility in responses to ensure both current users and potential customers of powerbanks could be considered when collecting data on interest in a new modular powerbank concept.
In regards to collecting information about the respondents themselves, the survey included questions about age and household income. Collecting age data, helps identify differences in preferences across different age groups and insight on who is willing to adopt the use of the modular powerbank. Asking about one’s income also indicates consumer behavior, based on one’s purchasing power and socioeconomic status. These are key indicators of whether they are willing to purchase this new product and help identify which income groups might be more receptive to it.
To introduce the modular powerbank, the survey presented brief educational information describing modularity and how it can be used with the new product. Other attributes of the powerbank are defined, to provide information for respondents to make the best possible choice of which portable powerbank they would prefer.
The survey included six attributes to characterize the modular power bank and compare it to existing products currently available in the market. These attributes were within the six choice questions for the respondent to answer, each containing three product alternatives with varying combinations of these product features.This structure helped identify which product features were most influential when the respondent was making their decision.
| Attribute | Description | Levels |
|---|---|---|
| Price ($): | Total cost of the battery pack in U.S. dollars | $40, $60, or $80 |
| Capacity (mAh): | How much charging power the power bank contains | 5,000 or 10,000 mAh |
| Charging Speed (mins): | Time required to charge a phone from 0% → 80% | 30, 45, or 90 minutes |
| Number of Modules: | Add-on modules that can be attached to expand capacity | 1, 2, or 3 modules |
| Weight per Module (g): | Added weight for each module | 150g or 250 g |
| Charging Method: | Compares wireless versus wired power bank, which helps determine what kind of charging method the respondent prefers | Wired only, wireless only, or both |
Using these six attributes allows for data to be collected in whether consumers would prefer the modular power banks and determine the importance of specific product features in this market. Based on some feedback from our pilot analysis, we made several refinements to improve clarity, realism, and data quality in our final conjoint survey. First, we updated the product images so they more accurately represent the modular portable battery pack described in our introduction and attribute definitions. The new visuals depict attachable modules, built-in cables, and durability differences to help respondents better visualize the product concept.
We reduced the total number of attributes shown in each choice task to make the survey easier and more focused for respondents. Based on some feedback, the number of attributes created some form of overload, making it harder for participants to evaluate all features at once. In the final version, we reduced the design to just six attributes instead of seven removing “Durability of modules” and focusing more on price, capacity, charging speed, charging method, and weight of modules. This reduction will improve data quality and respondent engagement.
Finally, we will include some more demographic and device usage questions to better understand participants’ tech usage behavior and their preferences. Some of the added questions may include “How often do you upgrade or replace your electronic devices?”, “Which devices do you most often charge with a portable battery pack?”, and “How often do you experience your device running out of battery during the day?” These additional variables will help us analyze similarities and differences in preferences across demographic and behavioral segments.
Data Analysis
Sample Description
Our final sample includes 94 respondents and 4,716 choice questions. Summary tables report the distributions of age, income, and technology upgrade behavior.
| Age group | Count | Percent |
|---|---|---|
| 18_24 | 13 | 17.1 |
| 25_34 | 17 | 22.4 |
| 35_44 | 21 | 27.6 |
| 45_54 | 15 | 19.7 |
| 55_64 | 5 | 6.6 |
| 65_plus | 5 | 6.6 |
| Income bracket | Count | Percent |
|---|---|---|
| 100_150 | 9 | 11.8 |
| 150_plus | 5 | 6.6 |
| 25_50 | 17 | 22.4 |
| 50_75 | 17 | 22.4 |
| 75_100 | 15 | 19.7 |
| na | 4 | 5.3 |
| u25 | 9 | 11.8 |
| Tech upgrade group | Count | Percent |
|---|---|---|
| Every 2-3 year | 33 | 41.8 |
| Every 4+ years | 14 | 17.7 |
| Every year | 6 | 7.6 |
| only when they stop working | 26 | 32.9 |
Data Cleaning
We initially recorded 342 survey responses in Supabase. Before estimating our discrete choice models, we implemented a series of data cleaning steps to remove incomplete or low-quality responses. All cleaning was conducted in R using timestamps variables and the CBC question fields.
We calculated both the total survey duration and the time spent specifically on the conjoint section using the recorded start and end times. We then restricted the sample to respondents who completed all six choice-based conjoints (CBC) questions by requiring non-missing answers for cbc_q1 through cbc_q6. This step removed respondents who dropped out before finishing the choice tasks. Next, we screened for any straight lining behavior. For each respondent, we checked whether they selected the same alternative number in every CBC question (for example, always choosing option 1 across all six tasks). Respondents exhibiting this pattern were flagged and removed from the analysis, as such behavior suggests low engagement with the trade-offs presented in the survey. Finally, we used the computed survey duration to identify anyone speeding through the survey. We excluded respondents who finished the survey in under 10 seconds. Completion times this short are unrealistic for a multi-question conjoint survey and are unlikely to reflect thoughtful consideration of the alternatives.
After applying these filters, we retained 262 acceptable respondents for our analysis and excluded 80 responses that did not meet our quality criteria. These data cleaning decisions are standard in discrete choice experiments and help ensure that our estimated preferences for the modular portable battery pack are based on respondents who meaningfully engaged with the survey rather than on incomplete, random, or unreliable responses.
Modeling
This estimated utility function for the portable power bank conjoint survey is represented as follows: \[ U_{ij} = 0.00210(\text{Price}_{ij}) + 0.00000782(\text{Capacity}_{ij}) - 0.01348(\text{Modules}_{ij}) + 0.00000246(\text{Weight}_{ij}) + 0.00158(\text{Speed}_{ij}) - 0.06183(\text{MethodWired}_{ij}) + 0.04099(\text{DurabilityStandard}_{ij}) + \varepsilon_{ij} \] The table below summarizes the estimated coefficients, standard errors, and statistical significance.
| Parameter | Estimate | Std. Error | z-value | p-value |
|---|---|---|---|---|
| price | 0.002102 | 0.001895 | 1.109 | 0.267 |
| capacity_mAh | 0.000008 | 0.000012 | 0.633 | 0.527 |
| modules | 0.013476 | 0.024529 | 0.549 | 0.583 |
| weight_g | -0.000002 | 0.000625 | -0.004 | 0.997 |
| speed_w | -0.001580 | 0.001728 | -0.915 | 0.360 |
| methodWired | 0.061832 | 0.062292 | 0.993 | 0.321 |
| durabilityStandard | -0.040993 | 0.062357 | -0.657 | 0.511 |
Results
WTP for each attribute
The figure above displays the estimated consumer willingness to pay (WTP) for each product attribute, aligned with the simulated 95% confidence intervals. Below is how each plot is interpreted:
Modules:
Within the modules plot, the WTP decreases slightly, as the number of modules increases. This suggests that consumers may view the modular components not as valuable in comparison to the other attributes.
Weight (g):
When it came to the weight attribute, the heavier the product, the less the consumer was willing to pay for the product. The wide confidence interval shows considerable uncertainty, reflecting how consumers prefer the lower weight.
Charging Speed:
Consumers had a positive WTP for faster charging speed. The slope was steeper than other attributes, indicating faster speed was generally valued.
Capacity (mAh):
Consumers seemed to not favor higher capacity of the battery, as there was a small decrease in the WTP. This may be due to the battery capacity not being the consumers main focus as a feature of the product, therefore this feature may have been overshadowed by other attributes such as the speed.
Charging Method (Wireless vs. Wired):
Wireless charging seemed to produce the largest jump in WTP among all the other attributes. The mean WTP was strongly positive, even though there was a wide range in the confidence interval. This attribute seems to be one of the more influential features for the consumer.
Overall, the key attributes driving consumer choice, seemed to be the charging method and the charging speed. The charging method had the largest WTP and the widest separation from zero. This makes it appear to be one of the most influential features in the consumer feature choice. When it came to the charging speed, faster charging had a positive impact on the WTP, indicating a strong preference for quicker performance.
Simulations
To evaluate how consumers would respond to realistic product configurations, we created a simulated market consisting of three alternatives, each of them representing competitive position in the market. The table below summarizes each alternatives used in the simulation.
| Alternative | Price ($) | Capacity (mAh) | Add-on modules | Weight (g) | Charge time 0–80% (min) | Charging method | Durability |
|---|---|---|---|---|---|---|---|
| Competitor A: Low-price fixed pack | 40 | 5,000 | 0 (fixed) | 150 | 65 | Wired | Standard |
| Competitor B: High-capacity fixed pack | 70 | 10,000 | 0 (fixed) | 250 | 65 | Wired | Standard |
| Our product: Modular pack | 55 | 10,000 | 2 snap-on | 200 | 45 | Wireless | Rugged |
Single Market Simulation
The single market simulation shows that all three products earn similar shares (around 30-35%). This shows a competitive balanced market. The modular pack achieved a predicted share of approximately 33% with a 95% CI between 29% and 38%. This is relatively balanced.
Multiple Market Simulation
The multi-scenario simulation shows that high performnance configurations (high capacity, fast charging) shift share toward competitor B.Low price configurations shift share towards competitor A. CIs for the alternatives overlaps, so that shows no single design dramatically outperformed the other.
Price and modularity have the highest potential even though the coefficients are not strongly significant. Since lower price create a small increase in market share, additional module give the product slightly higher predicted utility, and faster charging improves competitiveness modestly, larger changes in these attributes might produce clearer market advantages.
Opportunity for increasing demand:create a more aggressive performance differentiation. Higher charging speed, higher battery capacity, and wireless charging beyond the ranges tested could meaningfully shift performance.
Sensitivity Analysis
Of our tested attributes (weight, price, and speed), price and speed were the most impactful. When price is decreased, market share increases significantly. When price is increased, market share decreases. When comparing market share to price, price still has a small impact on market share. This means that when considering how to price a battery pack for maximized revenue, the battery pack should be priced high as this would not drive away consumers.
This sensitivity analysis is limited in that it compares one attribute at a time, ignoring any interactions that may be affecting our data. While it is information to consider in predictions, it is not necessarily the only factor in accurate decision making.
Final Recommendations and Conclusions
Conclusion
Our proposed product, a modular battery pack, does not have the potential to be competitive in the current battery pack market. In the willingness to pay analysis, as the number of battery packs increased, the consumers willingness to pay decreased slightly. This suggests that battery pack modules are not a bonus to consumers and would therefore not give our product a competitive edge. Based on our survey and analysis, we would recommend avoiding battery packs with many modules. Beyond this, we recommend fast charging and a wireless charging system, as these were the two most positively impactful attributes on consumers willingness to pay. The capacity of a battery pack had very little impact on WTP, however we would suggest keeping the weight low on any battery pack being introduced to the market. In the market, price is not a large factor in the amount of the market a battery pack product is able to take, so we suggest making an expensive light wireless battery pack.
Limitations
Looking ahead to the future, there is more important information to be collected to further our understanding of this market. We could include physical examples of modular battery packs to give our survey takers a better understanding of the product we are inquiring about. Since this is a possible product, this would add strength to our survey by limiting misunderstandings that may come from a written survey about a product that survey takers are unfamiliar with. We could also expand our understanding with direct comparisons between “our product” and other products in the survey to better understand where we stand in the market. There are several unknowns that may be affecting our findings. Our sample size was limited given the time frame and budget constraints. This impacts the reliability of our conclusions. Our sample population may also be affecting our results. Because we used a survey company (Prolific), our sample population is likely to be more familiar with technology or more tech-savvy. This creates the potential for sway in our survey results, since our product is a piece of technology.
Attributions
All members contributed equally throughout this project and report.
Appendix
