Start with one product people want, a budget you can afford to risk, and an order process that does not depend on you feeling well every day. Test demand before buying much stock. Calculate what each sale leaves after costs. Then build the store around the hours, physical work and customer contact you can realistically manage.
That is my starting point for anyone asking how to start an e-commerce business with chronic illness. Choosing the platform comes later.
I co-founded URevolution with Corinne Gray in 2018, three years after my diagnosis with chronic myeloid leukemia. Illness had introduced an unexpected problem into my life: people often did not know what to say. Conversations became awkward. Sometimes they disappeared altogether.
URevolution grew from wanting somewhere to talk honestly about disability and chronic illness. It began as a magazine and developed into a business combining publishing with clothing, gifts and other products. The connection was straightforward. People could read about experiences that resembled their own, and wear something that said what they believed.
What I can offer is a way to think through the decisions. Some are ordinary business decisions. Chronic illness changes their consequences: a complicated returns process can consume the afternoon you had available for everything else.
You can accidentally create a job with no sick leave, no guaranteed income and a customer-service inbox that follows you to bed. Flexibility has to be built into how the business works.
Work out what the business can ask of you
Before choosing products or paying for software, write a short description of the working life you need. Include medical appointments, care responsibilities and the energy that daily life already takes. Your business gets part of what remains.
Use your recent experience to estimate capacity. Your best week is useful information, but a poor basis for a permanent dispatch promise.
Consider three operating states:
| Your available capacity | What the business can do | What needs another arrangement |
|---|---|---|
| Enough for ordinary work | Fulfill orders, answer customers and develop the business within your planned hours | Keep some time uncommitted for problems |
| Reduced capacity | Handle existing commitments and urgent messages; suspend optional projects | Someone else covers time-sensitive tasks, or you reduce new orders |
| Unable to work | Your backup handles existing obligations, or the documented closure process starts | Stop new promises, contact affected customers and arrange refunds where needed |
This is an operating plan, not a medical assessment. Your own limits and clinical advice take priority over a schedule. If you experience post-exertional symptom worsening, a business plan should not become a reason to push through it.
Separate physical effort from concentration and emotional effort. Twenty minutes packing parcels, twenty minutes correcting bookkeeping and twenty minutes dealing with an angry customer may have very different costs for you.
Our guide to housework with fibromyalgia and ME/CFS offers a useful parallel: divide the work into tasks and consider the recovery cost, not simply whether you can get it done once.
Set an order limit before you need one
Suppose you have eight usable business hours in a week. You reserve three for customer support, bookkeeping and other administration, and two for interruptions. That leaves three hours for work attached to individual orders.
If making, checking and packing each order takes 15 minutes, the arithmetic gives you 12 orders. That is a planning ceiling under those assumptions, not a sales target or a recommendation for your health. If you cannot reliably work those hours, the ceiling falls.
Measure a trial order from beginning to end. Include printing a label, finding materials and getting the parcel collected. For downloads or outsourced fulfillment, measure the exceptions that still reach you.
Decide how you will limit sales: available stock, a small release, limited service slots or disabling purchases. Check the actual controls on your chosen platform. An order limit written in a notebook cannot stop checkout.
Choose a product and business model together
The best e-commerce business for someone with chronic illness depends on what work they can repeat, what help they have and what customers will pay for. A product can be easy to ship and exhausting to support.
| Model | Work it can reduce | Work and risks that remain |
|---|---|---|
| Digital products, such as templates or downloadable artwork | Storage, packing and physical delivery | Creating something useful, accessible files, compatibility questions, updates and finding customers |
| Print on demand | Buying finished stock in advance and packing each order yourself | Samples, design rights, product descriptions, supplier errors, customer support and potentially tight margins |
| Stock you pack yourself | Dependence on a third party for checking and dispatching each item | Storage, lifting, packing, collections, unsold stock and a backup for illness |
| Stock held by a fulfillment company | Much of the routine physical handling | Purchasing stock, storage and fulfillment charges, integration problems and returns arrangements |
| Handmade or customized products | The need to compete only on price or a generic design | Production time, revisions, deadlines and orders that may be difficult to delegate |
| Defined services sold online | Inventory and product manufacturing | Your availability, delivery deadlines, scope disputes and meetings unless the service is designed around asynchronous work |
A small downloadable product may suit someone who needs to avoid physical handling. A carefully tested print-on-demand offer may suit someone who designs but cannot pack. Neither is automatically passive income.
With dropshipping, another supplier ships a product you sell. You still need to understand its quality, delivery arrangements, safety obligations and returns process. A low entry cost can hide expensive customer problems. I would be particularly cautious about suppliers whose products you cannot inspect or whose support you cannot reach.
Subscriptions deserve separate scrutiny. Recurring revenue creates recurring promises. A monthly box or membership that needs new material every week can be a poor fit if nobody can cover an absence.
You do not have to sell your diagnosis
URevolution's purpose came directly from lived experience. That gave us a subject we had reason to care about and a point of view: disability belongs in diversity. Our publishing and products share that purpose.
Your business does not have to involve illness. You might sell sewing patterns, specialist parts, illustrations or something connected to work you did long before becoming ill. Choose an audience you understand and a useful offer you can deliver.
You also do not owe customers an ongoing account of your symptoms. A dispatch policy can say when an order will ship without explaining your medical history.
Find evidence that somebody will buy
Being encouraged by friends feels good. It does not tell you whether people will pay the price the business needs.
Start by describing one customer, one reason they might buy and one product. “Gifts for everyone” leaves you with too many decisions. “A downloadable planning tool for freelance editors managing several clients” gives you something specific to investigate.
Look at alternatives customers already use. Read reviews for recurring complaints. Ask potential buyers about their last purchase: what they bought, what it cost, what frustrated them and why they chose it. Those answers are more useful than asking whether they like your idea.
You can do this through short written conversations if calls are difficult. Ask permission before researching in a patient or disability community, follow its rules, and keep sales pressure out of spaces people use for support.
Then run a bounded test:
- Make one sample or prototype. For a physical product, inspect what the buyer would receive. For a download, ask someone to use it without you explaining every step.
- Show the real offer. Include the intended price, delivery details and limitations. A reaction to a free mockup is different from a buying decision.
- Choose the test limit in advance. Set a maximum spend, the work you can give it and the evidence that would justify continuing.
- Invite a small number of orders you can actually fulfill. If you are still exploring demand, a clearly described waitlist avoids taking money before you are ready.
- Review both the sale and the work. Did people pay? What questions did they ask? Could you fulfill the offer without sacrificing necessary care or recovery?
A few paid orders can teach you something. They do not prove the business can replace an income. Preorders also create customer obligations; do not treat customer money as risk-free development funding.
For URevolution, the underlying need was recognition and more honest conversation. The commercial question still has to be answered separately: does someone want this particular product, at this price, with these delivery terms? A worthwhile mission cannot answer that for them.
Calculate the money before building the store
Make two budgets: the money needed to start, and the money needed to keep operating. Keep your household essentials and healthcare costs visible alongside them, without treating those funds as available business capital.
The U.S. Small Business Administration's planning guidance distinguishes startup costs from ongoing expenses and explains break-even analysis. Australia's business.gov.au guide to starting a business also covers readiness, costs and cash-flow planning.
Your startup list might include samples, a domain, registration, necessary equipment and professional advice. Recurring costs can include the selling platform, apps, bookkeeping, insurance and paid support. Check the full price after any introductory offer.
Then calculate a sale. The figures below are invented USD examples, not URevolution's accounts, a supplier quote or an earnings forecast. Tax collected for government is excluded from revenue. Actual tax treatment needs local advice.
| One illustrative order | Amount |
|---|---|
| Product price paid by customer | $32.00 |
| Shipping paid by customer | $5.00 |
| Total revenue before tax | $37.00 |
| Product or production cost | −$15.00 |
| Actual shipping cost | −$6.00 |
| Packaging or fulfillment handling | −$1.00 |
| Selling and payment fees | −$1.50 |
| Allowance for refunds, faults and replacements | −$1.50 |
| Amount left before advertising and fixed costs | $12.00 |
| Advertising cost attributed to this order | −$5.00 |
| Amount left toward fixed costs and your pay | $7.00 |
That last figure is the contribution per order. It is not take-home pay.
At $140 in monthly fixed costs, you need 20 orders at a $7 contribution just to cover those fixed costs. To leave another $700 for yourself before personal tax, the same example needs 120 orders: ($140 + $700) ÷ $7. Your actual accounting treatment may differ with your business structure.
Now put that beside the capacity estimate. A business that needs 120 orders but can reliably handle 40 has a problem to solve before launch. You might need a different price, lower costs, a different product or a way to reduce the work per order.
Include your time even when you are not paying yourself yet. Record the total hours spent on selling, administration and support. If you outsource, put the helper's cost into the model. Free labor from someone you love is still labor with limits.
Check cash timing, not only profit
A payment processor may release money after your supplier wants payment. A refund may be due while you are still waiting for a supplier credit. Find out the actual timing for your providers.
Write down what must be paid before the next payout arrives. Include a realistic reserve for existing customer obligations. Test a month with low sales, a run of faulty orders, and a period when you cannot work.
If the model requires every order to go perfectly, revise it. If you need dependable income next month, a new store is a risky way to promise yourself that income.
Check benefits, registration and customer obligations early
Self-employment can affect disability benefits, income protection or other support. The rules depend on the scheme and where you live. Do not assume that a small turnover, no salary or leaving money in the business makes the work irrelevant.
For example, the U.S. Social Security Administration requires recipients to report relevant work and income changes, including starting or ending self-employment. Get advice about your particular benefit before trading; rules for one program should not be applied to another.
Take a short written description of your planned activity to the relevant benefits adviser or agency. Ask what to report, when to report it, which records to keep and how work activity is assessed. Keep the response.
You also need to establish the requirements for your business location, products and customer markets. Work through business registration and structure, taxes, insurance, product safety, intellectual-property permissions, privacy, marketing consent, shipping and returns. Use local official guidance and professional advice where needed. There is no single international checklist that settles all of these.
Selling only within a market you understand can make a pilot easier to manage. Shipping worldwide on day one adds decisions about duties, delivery and consumer rules before you know whether the core offer works.
In the United States, the FTC's merchandise-order guidance explains the need for a reasonable basis for shipping promises and the steps required when shipment is delayed, including seeking consent or refunding as applicable. A message announcing that you are ill does not, by itself, resolve an existing order.
Choose a platform you can operate on a difficult day
URevolution runs on Shopify. That is context for this guide, not proof that it is the right choice for your business.
Compare the work required as well as the subscription price. A marketplace may help people find a product, but it does not guarantee sales. It also sets rules you must follow. Your own store gives you more control over presentation and customer relationships while leaving you responsible for attracting visitors.
Hosted store platforms generally take care of some technical infrastructure. With a self-hosted store, establish who will maintain hosting, updates, security and recovery. The cheap option can become expensive if it creates work you cannot manage.
Before committing, trial these tasks with the input methods and assistive technology you actually use:
- Create and correct a product listing.
- Find an order and answer the customer.
- Change stock availability and dispatch information.
- Issue a refund and understand its consequences.
- Export useful records and give a helper appropriate access.
- Stop new purchases, then confirm that the change worked.
Test the admin area as carefully as the storefront. A site can be usable for customers while its management tools create barriers for you.
Our article on technology that supports disabled employees explains options such as dictation, alternative input devices and asynchronous working. Apply the same question to your own business: which specific task is difficult, and what would make it workable?
Start with the fewest tools that do the job. Each additional app brings another bill, setting or failure to understand. Compare current plan details and essential integrations before paying; this guide deliberately avoids prices that can go stale.
Build a small storefront that answers ordinary questions
Launch with a manageable selection. One well-explained product can teach you more than a sprawling catalog whose descriptions you cannot maintain.
Every product page should make clear what the customer receives, what it costs and when it is expected. Add the details that prevent avoidable questions: dimensions, materials, fit, care, compatibility or download format, as relevant. Show limitations plainly. Use photographs or previews that reflect the product.
For print on demand, distinguish production time from shipping time. For a digital product, say which software or device it requires and what the buyer is licensed to do with it. Do not promise lifetime updates casually.
Publish clear contact, shipping and returns information. Check that the policy fits both your legal obligations and what your supplier will actually do. A supplier refusing a replacement does not automatically end your responsibility to the customer.
Make buying accessible
Use Web Content Accessibility Guidelines (WCAG) 2.2 as a technical reference. Pay particular attention to the whole buying process, including product options, the basket, payment errors and order confirmation.
Can someone use the essential controls with a keyboard? Is focus visible? Are fields labeled and errors explained in text? Can a person enlarge the text and still use the page? Do images have useful alternatives, and do videos need captions? Can a customer get help without making a telephone call?
Digital products need their own accessibility checks. A store selling a PDF should consider the PDF's structure and reading order, not just whether the Buy button works.
Automated checks can help find defects. They cannot prove that the complete experience works for disabled customers. Budget for appropriate manual testing and, where possible, paid feedback from people using relevant assistive technologies.
Test an actual order from beginning to end
Use the platform's test facilities, then a small real transaction where appropriate to check the parts a simulation cannot prove. Follow the confirmation, supplier handoff, delivery, support and refund process. Check the records and costs too.
If someone else prints, packs or ships, ask what happens when the product is wrong, tracking stops or stock disappears. Know the contact route, evidence required and who pays for the remedy. Test the handover while you can still fix it without a queue of customers waiting.
Write the plan for when you cannot work
This is where much of the promised flexibility either becomes real or evaporates.
Keep a short operating document somewhere your authorized backup can access. Explain the immediate tasks in plain language. Record supplier contacts, current commitments and the route to the systems they need. Use individual accounts and appropriate permissions where supported; avoid putting passwords in the document.
| Task | What the backup needs to know |
|---|---|
| Existing orders | Where to find them, promised dispatch dates and the next action |
| Customer messages | Official inbox, response priorities and approved templates |
| Refunds and replacements | Who may authorize them, spending limits and the escalation route |
| New orders | Which controls stop purchases and how to verify each sales channel |
| Marketing | Where to pause ads, scheduled promotions and product launches |
| Payments and deadlines | Bills or filings requiring action, and the adviser to contact |
Rehearse one ordinary problem with the backup. A process that depends on you answering five follow-up questions has not yet transferred the work.
Co-founding URevolution is part of my story. It does not mean every reader has someone available to help. Nor should a partner become an unpaid emergency department for the business by default. Agree the task, availability, compensation where applicable and limits before relying on them.
If you work alone, the plan may be to stop taking orders earlier. You can also design the offer around limited releases, longer truthful dispatch windows or fewer time-sensitive services. Identify paid cover where the business can afford it, rather than assuming you must outsource everything immediately.
Know exactly what “pause” does
A banner saying you are taking a break may leave checkout running. A platform plan change may take effect later than expected. Closing one storefront may leave another sales channel accepting orders.
For example, Shopify's guidance on temporarily closing a store distinguishes restricting storefront access from its Pause and Build plan. It also warns that a plan change can have a future effective date and recurring app charges can continue.
Before relying on any pause feature, verify the effective date, each active sales channel, ongoing charges and how customers with existing orders can contact you. Complete, delay by agreement or refund those orders as required. Pause promotional messages that would invite purchases you cannot support.
Automate routine work, with a way to catch failures
Order acknowledgments, tracking updates, low-stock alerts and bookkeeping transfers are useful places to start. So are prepared replies for questions you answer repeatedly.
For each automation, decide what happens when it fails. A confirmation email must not invent a dispatch date. An AI support tool must not promise a refund, product feature or medical benefit you have not authorized. Give customers a clear route to a person, and check how any tool handles their data.
Choose one main customer-support channel and publish a response window you can meet. Acknowledgment and resolution are different promises. Turn off optional notifications outside working time.
Find customers without making constant visibility your job
Pick one realistic way for the right people to discover the product and one way to stay in touch with people who choose it. You can expand when there is evidence and capacity.
That might mean useful search content and an email list, a marketplace listing and repeat customers, or a focused partnership with a relevant organization. The work should fit the offer. A store selling replacement components has different marketing needs from an illustrator selling prints.
URevolution brings publishing and products together around disability and inclusion. You can see that relationship in the Embrace Diversity collection. It is an example of translating a point of view into a product, not a reason you need to build a magazine before selling anything.
For your first marketing experiment, specify the audience, the offer, the time or money you will spend and what you will measure. Start with product pages that answer buying questions. If you send marketing email, obtain the permissions required for your market and provide an appropriate way to unsubscribe.
Batch work when that suits you. Keep a small reserve of useful material rather than promising daily output. Scheduled campaigns need a pause procedure too.
Do not confuse a popular post with a viable business. Track sales, contribution after costs, refunds, support time and the work the campaign created. Avoid committing heavily to paid advertising until you understand the product's margin and know that checkout works.
Your health does not have to become content. You can be visible as a business owner without publishing proof that you are ill enough to need boundaries.
Launch in stages you can pause
A calendar can organize work. It should not turn a flare into a missed business milestone. Move through these stages when the evidence and your capacity allow.
| Stage | What to finish before moving on |
|---|---|
| Define the limits | Available work, maximum cash at risk, benefits questions and what happens if you stop |
| Test one offer | A usable sample, a real price and evidence from likely buyers |
| Check the economics | Costs, contribution per order, payment timing and an achievable order volume |
| Build the minimum store | Clear listings and policies, an accessible purchase route and a tested order process |
| Prepare for absence | Written procedures, realistic customer promises and a verified way to stop sales |
| Run a small pilot | Orders within your limit, recorded costs and an honest account of the workload |
| Decide what follows | Keep the offer, change it, pause it or close it based on the results |
During the pilot, keep a simple record of money received, variable costs, fixed costs and hours worked. Note which tasks caused problems or required help. A broad “good week / difficult week” note may be enough; you do not need to turn your medical life into another reporting obligation.
Review at a cadence you can maintain. Ask which products earn their place, which questions keep returning and which promises create pressure. Sometimes the right next move is to remove a product or stop a marketing channel.
Questions worth answering before the first sale
How much money do I need to start?
There is no reliable universal minimum. A simple download and a stocked physical-product business have different needs. Price your actual setup, one complete order, ongoing bills and a reserve for customer obligations. Set a maximum loss you can absorb without using money needed for essentials. A low monthly platform price is only one part of the budget.
Can I run the business without a co-founder?
Yes, but every task needs an honest owner or a decision to stop offering it when you are unavailable. Keep the first offer small. Use documentation and appropriate automation, investigate limited paid cover, and practice closing sales. A solo business should not depend on a person who has never agreed to help.
Should I tell customers about my illness?
That is your decision. Customers need accurate information about their order and how to contact you. They do not need your diagnosis to understand a dispatch window. If you want illness to be part of your public story, decide which details remain private and whether you want to answer follow-up questions.
What if the business makes my life harder?
Look at what is creating the strain: the product, order volume, deadlines, customer contact, costs or the business itself. Change what can be changed. You can also pause or close, while dealing properly with outstanding commitments. The money and effort already spent do not require you to keep spending them.
I wanted URevolution to make room for conversations that illness had made difficult. Building a business around that purpose does not make my circumstances a test anyone else has to pass.
Before you buy a domain or choose a theme, write down your first offer, who it is for, what one sale leaves after costs and what happens if you cannot work next week. If you can answer those questions, you have something concrete to test.
