E-learning: four design mistakes to avoid

Neglecting instructional design

An ob­serv­able ob­ject­ive helps you choose an activ­ity and a suc­cess cri­terion. Or­gan­ise con­tent around that task: an ex­pli­cit out­line and clear in­struc­tions are useful, but a short dur­a­tion or a story does not, on its own, guar­an­tee un­der­stand­ing.

To divide activ­it­ies between remote pre­par­a­tion and sup­por­ted prac­tice, see moving from classroom training to blended learning. The choice of format should start with the in­ten­ded task, not the number of an­im­a­tions.

Forgetting interactivity

A linear video can suit some ob­ject­ives, but op­por­tun­it­ies to prac­tise and ex­plan­at­ory feed­back can help learners check their un­der­stand­ing. Quizzes and scen­ari­os are useful only when aligned with the ob­ject­ives. Ex­plore this fur­ther in interactive online courses.

Underestimating the user experience

• Poor us­ab­il­ity: nav­ig­a­tion dis­cour­ages learners before they reach the con­tent.
• Over­looked ac­cess­ib­il­ity, in­clud­ing cap­tions, con­trast and screen read­ers: part of the audi­ence is ex­cluded.
• In­ad­equate tech­nic­al sup­port: bugs and slow per­form­ance un­der­mine ex­cel­lent con­tent.

Failing to measure and improve continuously

After pub­lic­a­tion, ex­am­ine in­dic­at­ors linked to the ob­ject­ives, col­lect learner feed­back and revise prob­lem­at­ic ele­ments. If you use a con­ver­sa­tion­al tool, check which data it col­lects and how those data can be ac­cessed before making it an ana­lyt­ic­al source. See the solu­tion for trainers.

Organise a review around the four existing mistakes

Use the art­icle's four themes as head­ings in a single audit, without adding others. For each, look for evid­ence in the module and clas­si­fy the risk: blocked access, in­ad­equately as­sessed learn­ing or an im­prove­ment to con­veni­ence. This hier­archy pre­vents a cos­met­ic cor­rec­tion from taking pri­or­ity over an am­bigu­ous in­struc­tion or an in­cor­rect pro­ced­ure. The review should lead to de­cisions about cor­rec­tions, not an over­all score that is dif­fi­cult to in­ter­pret.

Assign an owner and dead­line to each defect. A sub­ject-matter expert val­id­ates the con­tent; a de­sign­er checks align­ment between the ob­ject­ive and as­sess­ment; a member of the in­ten­ded audi­ence tests us­ab­il­ity. One person may hold sev­er­al roles, but the dis­tinc­tion between them must remain vis­ible. The de­liv­er­able is a re­gister con­tain­ing the loc­a­tion, prob­lem, con­sequence, pro­posed cor­rec­tion and val­id­a­tion status.

Correct design and interactivity together

For the first two mis­takes, start with the final as­sess­ment and work back­wards to­wards the con­tent. If the skill in­volves choos­ing an action, check that the ex­er­cise does not merely ask learners to re­cog­nise a word. If it in­volves per­form­ing a phys­ic­al action, a writ­ten ques­tion cannot re­place ob­ser­va­tion of that action. Remove ele­ments that do not pre­pare learners for any ex­pec­ted action, or place them in a clearly op­tion­al fur­ther-learn­ing re­source.

Next, review the in­cor­rect an­swers. A plaus­ible error re­veals a con­fu­sion to ad­dress; an absurd answer makes a high score easier without re­veal­ing un­der­stand­ing. For every dis­tract­or re­tained, write feed­back ex­plain­ing the over­looked cri­terion. Avoid giving away the answer in the word­ing of the ques­tion itself. Also have a com­pet­ent person try the module: if they fail be­cause of vocab­u­lary or in­ter­face issues, the as­sess­ment needs re­view­ing.

The de­cision rule is simple: if the activ­ity does not dis­tin­guish cor­rect un­der­stand­ing from a lucky answer, change it before adding visual ef­fects. Keep the old ver­sion so you can ex­plain what changed and why.

Test the experience beyond its appearance

For the third mis­take, W3C WAI re­com­mends plan­ning media ac­cess­ib­il­ity from the outset and dis­tin­guishes, among other ele­ments, cap­tions, tran­scripts, visual de­scrip­tion and an ac­cess­ible player. Its re­source, up­dated in Septem­ber 2024, provides tech­nic­al guid­ance; it does not allow a pro­gramme to be de­clared com­pli­ant without ex­amin­ing its im­ple­ment­a­tion. A cap­tioned video there­fore does not re­solve every access issue.

Build ac­cept­ance-test­ing tasks: find an in­struc­tion, pause, resume after an in­ter­rup­tion, answer an ex­er­cise and access sup­port. Record the device, browser and as­sist­ive tools used. Test real con­di­tions: noise, no head­set, lim­ited avail­ab­il­ity or work­place com­puter re­stric­tions. A suc­cess­ful test on the de­sign­er's com­puter does not es­tab­lish that the audi­ence will have the same re­sources.

• Clas­si­fy an obstacle as a block­er when it pre­vents an in­ten­ded learner from ac­cess­ing an es­sen­tial activ­ity.
• Provide an equi­val­ent al­tern­at­ive when voice or a par­tic­u­lar in­ter­ac­tion cannot be used.
• Dis­tin­guish ed­it­or­i­al check­ing of cap­tions from a broad­er tech­nic­al audit, and name the person re­spons­ible for each.

Close the measurement and correction loop

For the fourth mis­take, Samira pre­pares an ob­ser­va­tion sheet for a sim­u­lated count: check­ing the item ref­er­ence, hand­ling a dis­crep­ancy and jus­ti­fy­ing the de­cision. She keeps tech­nic­al help re­quests sep­ar­ate. After the first re­lease, she ex­am­ines re­cur­ring errors and in­vest­ig­ates their origin before making cor­rec­tions. A wrong answer may stem from a poorly ex­plained pro­ced­ure, a miss­ing pre­requis­ite or an am­bigu­ous case; the remedy dif­fers.

The mon­it­or­ing de­liv­er­able in­cludes a pri­or­ity change, its reason, its date and the ex­pec­ted out­come of the next check. If a cor­rec­tion af­fects a work­place rule, the expert re­val­id­ates it. If it af­fects only a button label, tar­geted ac­cept­ance test­ing may suf­fice. This ap­proach keeps the four mis­takes as a stable frame­work while evid­ence and cor­rec­tions evolve across ver­sions.

Key takeaways

These four themes provide a review frame­work, not a guar­an­tee of res­ults. Design around the ob­ject­ives, make activ­it­ies ac­cess­ible, test the pro­gramme and im­prove it using ob­ser­va­tions.