Social Media Strategy
Deciding which platforms deserve your attention, and which do not
Most social strategies fail at the first decision. Being present on five platforms badly loses to one platform run properly, so the useful half of a strategy is the list of places you have decided not to be, and the reason why.
What we usually inherit
Four ways a social strategy goes wrong before anything is posted
Almost none of these are content problems. They are decisions that were never made, or were made once by someone who has since left.
- The platform mix was never chosen, only accumulated.
- Someone opened an account during a launch, someone else opened one because a competitor had one, and a third was created for a campaign that ended. Nothing was ever closed. The result is a portfolio nobody selected, spread across teams, splitting attention that was barely enough for one platform.
- The format was chosen before the capability existed.
- A plan that depends on three vertical videos a week is a plan that depends on someone filming, editing and appearing on camera every week for a year. If that person does not exist and has not been hired, the plan will last about six weeks. Choosing the platform whose native format matches what you can genuinely produce is more important than choosing the platform with the biggest audience.
- Content pillars are topics, not commitments.
- A pillar called "industry insights" tells nobody what to do on Monday. A pillar becomes useful when it names who produces it, in what format, how often, and what the first three pieces are. Without that, the pillars sit in a deck and the actual output reverts to whatever is easiest that day.
- Success is defined with a metric borrowed from paid media.
- Judging a verification channel on last-click enquiries is like judging a showroom on how many people walk in and buy without speaking to anyone. The measure has to match the job. If the job is to make a buyer more comfortable proceeding, the evidence for that lives in branded search, sales conversations and time-to-close, not in the post-level dashboard.
Choosing where not to be
What each platform is genuinely good at, and what it costs you
This is the table we build for your business specifically. The pattern below is the general shape of it: every platform has a real strength, a real weekly cost, and a set of conditions under which the right answer is to leave it alone.
| Dimension | What it is genuinely good at | What it costs you each week | When to leave it alone |
|---|---|---|---|
| Reaching people by job title, and giving a business buyer something to read before a call. Individual accounts carry further than company pages. | Someone senior writing, or being interviewed, every week. It cannot be handed to a junior who has never sat in a sales conversation. | Your customers do not buy at work, or nobody in the business will put their own name to a point of view. | |
| Showing work that photographs well, and letting someone check that you are real before they book. Strong for hospitality, retail, clinics and trades with visible results. | A continuous supply of original photography and short vertical video. Stock imagery is recognised as stock imagery within one scroll. | The product is invisible, the results are confidential, or nobody on site will pick up a camera. | |
| TikTok | Reaching people who have never heard of you, because distribution follows the content rather than your following. | Volume, fast iteration and a tolerance for most things not working. It is a production commitment rather than a posting one. | You can produce one polished asset a month rather than several rough ones a week, or your approval process cannot move at that speed. |
| YouTube | Answering the questions people ask before they buy, and earning search visibility that keeps working long after publication. | Real scripting and production time per video. The payback period is long and the archive, not the individual video, is the asset. | You cannot commit beyond six months. A channel abandoned at video nine is worse than no channel. |
| Local community reach, groups, events, and an audience that skews older than the platforms that get discussed more. | Very little, if it is treated as a community and listings surface rather than as a content channel. | You are using it as a broadcast feed for a business page and wondering why organic reach is close to nothing. |
How the decision gets made
From guesswork to a shortlist with reasons attached
Establish how your customers actually buy
Interviews with your sales team, a read of recent won and lost deals, and the questions that come up before someone commits. This tells us whether social has a discovery job, a verification job, or neither.
You get: A short written account of the buying path
Audit every account, including the ones you forgot
Active, dormant, unofficial and duplicate. We look at what a prospective customer sees when they check you, which is rarely what the marketing team assumes they see.
You get: Account inventory with a recommendation on each
Score platforms against your business, not the market
Audience fit, format fit, production capability, competitive space and the cost of being average there. A platform can be enormous and still be wrong for you.
You get: A ranked shortlist with the reasoning written down
Decide what to stop
The uncomfortable half. Which accounts close, which are archived, which stay as maintained profiles with correct information and no pretence of activity. Getting this signed off is what frees the capacity for everything else.
You get: A decommission plan with owners and dates
Write pillars as commitments
Each pillar names an owner, a format, a cadence and its first three pieces. If we cannot write those three examples now, the pillar is not real and it comes out.
You get: Content pillars with worked examples
Agree the measure and the review date
What evidence would persuade you the channel is worth continuing, and when we will look at it. Both are settled before publishing starts, not after the results are in.
You get: Measurement framework and a fixed review date
Measurement
How to judge a channel that is not direct response
When social is doing a demand job, judge it the way you judge any acquisition channel: enquiries, qualified enquiries, cost per acquired customer. That is the straightforward case and it is less common than people expect.
When social is doing a verification job, the honest answer is that no single number will settle it. Nobody sees a post, screenshots it, and arrives at your enquiry form. They check you, feel better or worse about proceeding, and then convert through a channel that takes the credit.
So we triangulate. Branded search trend, whether the profile was visited in the days before an enquiry, what prospects mention unprompted on calls, the ratio of saves and shares to passive likes, and the quality of inbound rather than the volume. None of these is proof on its own. Together they are usually enough to make a budget decision without pretending to a precision that does not exist.
The one measure we will not accept as the goal is follower count. It is the easiest number to move, the easiest to buy, and the one least connected to whether anybody bought anything.
Questions
What people ask when choosing platforms
How many platforms should we actually be on?
For most businesses, one or two done properly, with a third only if a different audience genuinely lives there. The constraint is not budget, it is how much original material your business can produce in a week without the quality dropping.
If you can sustain one strong platform and a well-maintained profile everywhere else, that is a better position than five mediocre feeds. The mediocre feeds are visible to exactly the people you are trying to impress.
What happens to the accounts we stop using?
They get closed, archived, or converted into a maintained profile that carries correct information and nothing else. What they do not do is sit there with a post from two years ago at the top.
This is usually the fastest improvement available. It costs a morning and it removes the impression that the business lost interest halfway through something.
Our competitors post every day. Does that matter?
Only if it is working for them, which is worth checking rather than assuming. Frequency is the easiest thing to observe from outside and one of the least informative.
What we look at instead is whether their content could only have come from their business. Volume is imitable in a week. A genuine point of view is not.
How do you choose content pillars?
From two directions at once. What your buyers need to see before they commit, and what your business can produce repeatedly without inventing anything. A pillar that depends on material nobody can supply is a wish.
Each pillar gets a named owner, a format, a cadence and an example. If we cannot write the first three examples during the strategy work, the pillar does not go in the document.
How long before we know if the strategy is working?
Set the review date at the start, usually two quarters out, and agree in advance what evidence would count. Deciding that afterwards is how social budgets get defended or killed on the basis of whichever number looked best.
Some signals move sooner. Branded search, profile visits before an enquiry, and what prospects mention unprompted on sales calls tend to shift before anything in the platform dashboard does.
Get a straight answer on where your social effort should go
We will look at how your customers buy, what you can realistically produce, and what your current accounts are saying about you. Sometimes the recommendation is to do less in more places than you expect.
Related
Where to go next
- how we approach social overallThe pillar this strategy work sits inside.
- running the channel month to monthWhat happens once the strategy is agreed.
- putting budget behind socialWhen organic reach is not enough on its own.
- the messaging the channel repeatsPillars fail when the messaging underneath them is vague.
- wider content strategy
- who should own the workResourcing is a strategy decision, not an afterthought.
Last updated · Reviewed by Zubair Afzal