Strategy
Part of A social media training strategy that starts with the job, not the course catalogue
Six social media training strategy mistakes, from buying content early to live-account practice
Six social media training strategy mistakes in England, from undefined tasks to shared logins, with NCSC and GOV.UK testing advice and a fix for each.
The costliest social media training strategy mistakes happen before delivery: an undefined job task, an invalid measure, a mismatched route or a safety gate treated as optional. This unranked list covers six planning failures and gives a fix for each one.
What to take away
- Define the job task and the observable action before you buy content or pick a provider.
- Attendance and module completion are delivery facts, not proof of competence at work.
- Keep one synthetic task and rubric when you compare training routes.
- Rehearse with synthetic identities and get security sign-off before any live posting.
- Report observed change honestly until the evaluation design supports a causal claim.
1. Buying content before defining the task
A broad promise to "improve social media" gives neither a provider nor an assessor a stable target. Start with a recent workplace artefact, such as a post that went out without approval, and write down the action the learner should have taken.
Then test whether authority, workload or account access is the real cause. If a missing approval owner explains the failure, a course will not fix it. Fix the rota first, and buy training only for the gap that remains.
2. Reporting attendance as competence
Invitations, attendance and module completion are delivery facts. Competence needs a defined task, stated conditions, a rubric and a defensible judgement by someone independent of the trainer. Our selection checklist lists the evidence a course should be able to show.
A short private course cannot claim regulated standing by borrowing apprenticeship language. If a provider describes a qualification as regulated, check the exact qualification and awarding organisation on Ofqual's Register of Regulated Qualifications; the Regulated Qualifications Framework (RQF) is a framework, not a certificate. For apprenticeship funding claims, check the Education and Skills Funding Agency (ESFA) apprenticeship funding rules.
A course can, however, stop using completion as a stand-in for ability. Report completion and competence as two separate numbers, each with its own denominator.
3. Changing the test between channels
Comparing a live workshop with self-directed learning means little if learners get different briefs, support, time or marking. Lock one synthetic task and one rubric. Record access adjustments openly instead of quietly lowering hidden criteria.
Accessibility is best tested regularly using both automated and manual testing (see the accessibility testing page). That is a public-service standard, not a course certificate, but it shows why one automated check is not enough.
For public-sector websites and apps, the Public Sector Bodies (Websites and Mobile Applications) (No. 2) Accessibility Regulations 2018 require an accessibility statement. WCAG 2.2 is a W3C standard you can use as a testing reference; do not describe either as a social-media training certificate. Our product comparison method uses one common unit.
4. Practising in an uncontrolled live account
Shared credentials and unapproved posting turn a learning exercise into a possible incident. With 2-step verification you have a way of double checking that you are the person you claim to be when you use services such as social media (social media safety guidance).
Rehearse with synthetic identities and assets. Define how a learner's access is removed at the end. Let the security owner approve any later live practice in writing. Our risk scenarios show how this failure plays out.
5. Combining unrelated approvals
For personal data, the UK GDPR and Data Protection Act 2018 are relevant, with the Information Commissioner's Office (ICO) as the data protection regulator. You must determine your lawful basis under the UK GDPR before you start using personal information, and document it.
Give privacy, electronic marketing, advertising, security, accessibility, education status and procurement separate owners. For electronic marketing, check the Privacy and Electronic Communications Regulations (PECR); the CAP Code, administered by the Advertising Standards Authority (ASA), applies to non-broadcast advertising claims. A pass from one owner cannot stand in for another. See learner data protection for the privacy gate in detail.
6. Claiming business impact from a before-and-after chart
A metric can improve after training because of seasonality, staffing, campaign changes or measurement error. Before you claim impact, think about what matters when designing an evaluation. Without that design, a before-and-after chart cannot show that the training caused the change.
Report observed change honestly until the design supports a causal claim. Our return on investment method separates activity, competence and outcomes. Keep a short decision log: task, population, assessment, routes, gate evidence, pilot cap, stop rule and decision date.
Quick reference
Mistake / Fix
- Buying content first
- Write the task and the observable action first
- Attendance as competence
- Report completion and competence separately
- Different tests per channel
- One synthetic task and one rubric
- Live-account practice
- Synthetic identities, written security approval
- One combined approval
- Separate owners for each gate
- Before-and-after chart as proof
- Say "observed", and plan the evaluation design
Before you act
- Define the task, conditions and rubric before the pilot.
- Separate privacy, security, advertising, accessibility and procurement decisions.
- Lock one task and rubric across compared routes.
- Use synthetic identities and assets for practice.
- Cap the pilot and set a stop rule.
- Record the decision date and named reviewers.
