Maintenance
Part of Digital products ranked: which ones earn a place
Digital Products Mistakes After Launch: What to Fix First
Digital products mistakes found after launch, matched to the symptom you can see and put in a repair order: rights first, then refunds, support and review.
The product is live and something is off. Each of these mistakes is a decision made early, for a sensible-sounding reason, that only reveals its cost after the thing exists. None of them is a failure of craft.
The order of repair matters. Some mistakes compound if left, and some only cost money. Start from what you can see, work back to the cause, and fix the one that can force a rebuild before the one that merely irritates.
What to take away
- Rights problems found after launch are the only ones on this list that can force a rebuild, so they go first.
- Write the refund policy next. It stops you making the same decision repeatedly under pressure.
- A product with no written aftercare is a job you gave yourself without a salary.
- Refund rate is feedback on the sales page at least as often as on the product.
Match the symptom to the mistake
Start with what you can see. Each symptom below points at one mistake from the nine, and the first move is usually small.
| What you see | Likely mistake | First move |
|---|---|---|
| The second refund request is answered differently from the first | No refund policy, then an improvised one | Write the policy in plain words where a buyer sees it before purchase |
| Refund reasons cluster around "not what I expected" | A sales page that told buyers something different | Edit the page, not the product |
| Cancelling is awkward and reviews mention it | A subscription treated as a product with a payment attached | Describe the exit plainly and fix what would embarrass you |
| Questions, access problems and corrections keep arriving | An unbudgeted support tail | Decide the support window and the update commitment |
| The product describes a subject that has moved on | No review date | Update it, mark it as historical, or take it down |
| Somebody points at a font, image or code you included | Material you had not checked you could sell | Check every included asset |
Why the repair comes late
Eight of the nine mistakes are decided before launch, and each shows its cost at a different moment. The table gives the moment for all nine.
Mistake / The moment it costs you
- Building first
- Launch week, to an empty room
- Over-scoping
- The month you cannot finish
- Borrowed pricing
- Slowly, as the line underperforms
- Unchecked assets
- Whenever somebody notices
- Improvised refunds
- The second request
- Undesigned cancellation
- The first bad review
- Unbudgeted support
- Month three, every month after
- No review date
- A year later, quietly
- Misread refunds
- Never visibly, which is the problem
When each mistake costs you
- Launch weekBuilding first
- The month you cannot finishOver-scoping
- SlowlyBorrowed pricing
- Whenever somebody noticesUnchecked assets
- The second requestImprovised refunds
- The first bad reviewUndesigned cancellation
- Month three, every month afterUnbudgeted support
- A year later, quietlyNo review date
Borrowed pricing is the slow one. You cannot see other creators' costs, their audience, their support burden or whether their price is working, so what looks like a market rate is a set of guesses that happen to be visible. See why published figures fail before you reprice against them.
Fix first: the assets you cannot prove you may sell
Fonts, stock images, sample code, music, quoted passages, a template somebody sent you, work a collaborator made. Personal-use licenses commonly forbid resale, and a permission that existed when you downloaded something is not a record of what it permits now.
Check every included asset. The United States Copyright Office's registration toolkit is worth reading if the product is substantial enough that you would want to defend it. This is the only mistake here that can force you to rebuild.
Fix second: refunds, read for information
Write the policy before the next request, in plain words, where a buyer sees it before purchase. It protects the buyer, and it protects you from making the decision twenty times.
Then read the refunds. They carry information, and the information is often about the sales page rather than the contents, as affiliate income metrics show for another kind of figure. A buyer who expected something different was told something different. Ask each person what they expected, and where the answers cluster, edit the page.
Claims about results must be ones you can support. The FTC's endorsement and testimonial guidance says an advertiser needs adequate substantiation for results claims, and it applies to your page as much as to a sponsor's copy. Keep refund records too; creator taxes questions covers what to record.
Fix third: the exit and the support tail
Recurring pricing changes the product. It commits you to a supply of new value, and it commits you to a cancellation experience that a buyer will judge you on. Design the exit before the entrance.
The Federal Trade Commission's negative option rule page sets out expectations around automatic renewals in the United States, and a cancellation path that would embarrass you if described plainly is a business risk regardless of jurisdiction. The membership mistakes operators repeat are worth a read, as is the note on recurring income.
Support is the cost that most reliably goes unbudgeted. Questions, access problems, corrections, a subject that changed under the product, and the buyer who is sure something is broken when it is not are the ordinary condition of having sold something. They continue as long as the product is on sale, so set the support window and update commitment.
Retire, update or give away
Anything that references a changing subject ages, sometimes fast. A product still listed a year after it stopped being accurate does more damage than a product that was never made.
Set a review date and put it in the calendar. Three outcomes are allowed on that date: update it, mark it clearly as a historical document, or take it down. A useful test is whether you would recommend the product to a friend without a caveat, and that test is more reliable than sales, which lag the decision.
Common questions
I have already launched and several of these apply. What first?
Rights, then refunds. Rights because it is the one that can compound, refunds because a written policy stops you making the same decision repeatedly under pressure.
Is a small audience a reason not to build a product?
No, it is a reason to build a smaller one and to ask people directly what they need. Small and specific beats large and vague at this size.
How do I know when a product should be retired?
When you would not recommend it to a friend without a caveat. That test is more reliable than sales, which lag the decision.
Should I ever give the product away to fix a complaint?
Often yes, quickly, and without negotiating. The time you would spend defending a small sum is worth more than the sum, and the goodwill is worth more than both.
