Algeria AgencyDigital marketing
Six months of posting, nothing: the diagnosis in order
Five possible causes, one that everybody tests first, and it is fourth by likelihood. The order is the method.
The article accompanying this one describes what you buy when you buy this service: a quantity of posts rather than an effect, and a set of things the supplier cannot know for you. It is written for the moment of deciding.
This page is written for six months later, when the question becomes: we have been posting regularly for six months and nothing has happened — what is wrong?
It is the most frequent question in this area and it is almost always handled badly, for a simple reason: everybody tests the same cause first — the content — because it is the only one they can see. It is fourth by likelihood, and testing it first costs three months.
This article gives the order. It starts with a question that is not a cause at all, then works through the five possible causes from the most frequent and cheapest to correct to the most expensive, and ends on the case where the right decision is to stop.
First of all: is it already working?
The first step is not a cause, it is a check, and it resolves a non-trivial share of cases. Many businesses conclude nothing is happening when something is happening and is not being counted.
The mechanism is ordinary. Results from this channel rarely arrive as a measured click: they arrive as a private message, a call, somebody introducing themselves with "I saw you", or an existing customer ordering more because you are on their mind. None of those paths appears in the platform’s statistics.
The check takes two weeks and consists of asking everybody who contacts you: how did you hear of us? Note the answer in one line. It is crude, it is sufficient, and it settles half the situations.
A second check is worth ten minutes: look at your sales or enquiries over twelve months, not over the last quarter. Slow growth is invisible month by month and obvious over a year, and many people conclude failure from too short a graph.
If those two checks show nothing, then genuinely nothing is happening and the rest of this article applies. If they show something, the problem was not the channel but the measurement — better news than it sounds, because what is measured can be increased.
Cause 1: it is not your customers’ platform
This is the most frequent cause and the least often considered, because the platform choice is made at the start, in five minutes, and never revisited.
The choice is generally made on what the owner uses or on what is fashionable, and both criteria are unrelated to where the customers are. A business selling to tradespeople and posting where students are will produce sound content that reaches nobody useful.
The check needs no tool: ask ten existing customers where they spend their time and what they use to find a supplier. Ten answers are enough to settle it, and they often contradict intuition — particularly for businesses selling to other businesses rather than to individuals.
One local case is worth flagging because it is frequent and costly: many businesses post on a visual platform while their entire customer relationship happens in a messaging app. They produce images for one audience and receive their enquiries somewhere else, with no link between the two.
If this is the cause, the correction is immediate and free — change platform — which is why it is tested first. Six weeks is enough to see whether the same content gets a different response elsewhere.
Cause 2: nobody replies fast enough
The second cause is the one the supplier cannot fix, which is why it survives every change of agency.
The mechanism is described in the neighbouring article and is worth repeating because it is the most invisible cause of failure: on these channels the person who saw your post writes, and they write to three businesses. The first to reply gets the conversation; the others receive a silence they read as disinterest.
The check is mechanical. Take twenty messages received last month and calculate the delay between arrival and your first reply. If the median exceeds a few hours, you have found the cause and there is no point looking further.
The trap is that this delay is invisible from inside. Nobody says "we reply in eight hours"; they say "we always reply". The calculation over twenty real messages consistently returns a higher figure than the estimate, often by a factor of three.
The correction is not marketing: it is an organisational decision about who watches, when, and with what authority to reply without asking. Until it is taken, increasing the number of posts only increases the number of lost conversations.
Cause 3: there is nothing to do after seeing
The third cause is structural and frequent among businesses that post well: the content is seen, appreciated, and leads nowhere because no action is offered.
The symptom is recognisable: decent reactions, positive comments, and zero enquiries. It is precisely the case that makes people say "engagement is worthless", when the problem is not the engagement but the absence of a next step.
What is missing is almost always one of three things. A reason to write now rather than later. A price or an order of magnitude, whose absence makes people give up before asking. Or simply a mention of what you sell, which disappears from posts when people try not to look commercial.
The correction is the cheapest of all and the fastest to test: add to one post in three a sentence saying what to do and how to reach you. Not to all of them — a page that sells in every post stops being read — but to one in three, which is enough and shows within a fortnight.
Check the path itself too. If your post points to a profile whose description does not say what you do, or to a number that does not answer messages, the action exists on paper and not in fact.
Cause 4: the content, at last
This is the cause everybody tests first and it appears here, fourth, for a reason that is not ideological: the three before it are more frequent, cheaper to correct, and faster to test.
When it genuinely is the cause, it almost always takes one of two forms. The first is generic content: quotations, public holidays, advice applicable to any business. It does no harm and produces nothing, because nothing in it is specific to you.
The second is content that talks about the business rather than to the customer. "Our team is growing", "we attended this trade fair", "happy holidays to all": those posts interest the business and nobody else, and they occupy the space of what would have worked.
What works is dull to produce and reduces to one sentence: show the work. A repair in progress, a product before and after, a customer problem solved, a frequent question answered. It requires no creativity and never runs out, because the material is produced by the activity itself.
The test runs over six weeks and on one change: replace generic content with work content, without touching the rhythm or the platform. It is the only way to know whether that really was the cause.
Cause 5: the volume is too low to conclude
The fifth cause is not a defect but an error of judgement, and it explains a share of premature abandonments.
Three posts a month for six months is eighteen posts. That is too small a sample to conclude anything from, and the absence of a result is not information — it is the expected result, including if everything else is correct.
The threshold below which you should not conclude is not a universal figure, but the order of magnitude is stable: below one post a week sustained for several months, a social channel has had no opportunity to produce anything, and the diagnosis should be suspended rather than delivered.
This cause has one feature that makes it painful: the correction requires doing more of the thing you suspect is not working. That is uncomfortable, and it is nonetheless the only way to get an answer — or to decide knowingly to stop, which is section 10.
One useful clarification: the volume in question is of posts, not of effort. A business producing one very worked post a month is in this situation, and the correction is not to work harder but to post more often with less care per piece.
The order, and why it matters
The five causes are not tested in any order, and the one proposed here follows two combined criteria: frequency, and the cost of the test.
The platform is tested first because it is the most frequent cause and the test is free — repost the same content elsewhere for six weeks. Reply time comes next because the calculation takes an hour and the correction costs nothing either.
The missing action is third because it is corrected with one sentence. Content is fourth because changing it costs production. And volume is fifth because raising the rhythm costs the most, in time and in consistency.
This order has an important practical consequence: only one cause is tested at a time. Changing the platform and the content in the same month gives a result nobody can attribute, and you arrive six weeks later with no more information than before, having spent the same effort.
Record each test and its date. A business that has tested three causes in eighteen months knows something solid about its own market; a business that has changed everything three times knows nothing, and that is the commonest situation.
What not to change first
Three changes are systematically proposed in this situation and none of them is a diagnosis. They are expensive and they move attention.
The first is redoing the visual identity. It is enjoyable, it produces a feeling of renewal, and it has no effect on the five causes above. A business whose posts nobody sees will not have them seen more because the colours changed.
The second is changing supplier. Sometimes justified, but rarely diagnostic: if the cause is reply time or a missing action, it will survive the change intact, and the business will conclude after another twelve months that this channel does not work for it.
The third is adding a platform. It is the opposite of what is needed: the maintenance load doubles, quality falls on both, and the original cause goes untreated. The question "should we add one" only arises once a platform is working.
One exception deserves flagging: if the diagnosis establishes the platform as the cause, changing it is diagnostic — but that is a replacement, not an addition, and section 2 is what says so.
The six-week test
Every cause is tested the same way, and standardising the test is what makes the results comparable with each other.
Six weeks, one change, nothing else modified. Six because it is long enough for chance to wash out and short enough not to spend the year on it; one change because two make the result uninterpretable.
Two figures are recorded before and after, and they are not the platform’s: the number of conversations started, and the number of enquiries mentioning having seen you. Views, reactions and followers are of no use to this diagnosis, because they can move a great deal without anything changing commercially.
Record both figures before starting. It is the part everybody skips and without which the test concludes nothing: after six weeks nobody remembers the previous state, and impression replaces measurement.
At the end, one decision: the change produced something and is kept, or it produced nothing and is reversed before testing the next cause. Reversing matters — stacked changes produce a configuration nobody chose.
When the conclusion is to stop
This outcome has to be planned for because it is legitimate and because nobody ever proposes it: after testing the causes, it sometimes turns out that this channel does not suit this business.
The typical case is a business whose customers do not search that way: an industrial supplier whose buyers go through tenders, a trade where you are recommended by word of mouth, a very local business whose market fits in a neighbourhood. Such businesses can post competently for two years with no result, not through an execution error but a channel error.
Stopping is not a failure, and it is often the decision that unblocks everything else: the effort and budget move to what works for that trade — direct relationships, presence in a trade directory, search, physical signage.
There is however one thing not to do when stopping: leave the accounts as they are. An abandoned account whose last post is eight months old is worse than no account, because it is found and read as a business that has ceased trading.
A clean exit means updating the contact information, publishing one last permanent thing — what you do, how to reach you — and leaving it there. It becomes a business card rather than an abandoned site.
What six months should have produced as information
Even when nothing worked commercially, six months of posting produces information with value, and almost nobody recovers it before throwing everything away.
The first is the list of what was looked at most. Three posts out of forty did clearly better than the rest; knowing which and why is worth more than the overall average, and it is the material for the next format.
The second is the list of questions asked in comments or messages. It is a free inventory of what your customers do not understand, and it feeds directly into your site’s pages — the ones the article on owned content describes.
The third is the time actually spent. A business that sustained six months now knows what a post costs it, information it did not have at the start and which decides what is sustainable next.
Recover those three before changing anything. A business starting over after six months begins having lost everything; a business that records those three lists begins with a format, a subject and a realistic time budget.
What we do, and what we refuse
What we do is bounded. We run the diagnosis in order, we calculate the real reply time on your messages, we run the six-week tests one at a time, and we recover the three lists in the previous section before any change.
We refuse to change more than one thing per test. It is the most frequent request when things are stuck — "let us change everything and see" — and it guarantees that after six weeks nobody will know what produced the result or how to reproduce it. A test that teaches nothing costs the same time as a test that teaches.
We also refuse to maintain a monthly retainer on a channel the diagnosis says to stop. It is a commercially expensive position and the only tenable one: billing for the production of posts we know have no effect is the opposite of the work we claim to do.
And we do not promise that this channel will work for your business. For a proportion of trades it does not, whatever the quality of execution, and saying so after three months beats absorbing it for two years. There is also a check you can make this week without us: ask ten existing customers how they found you.
Frequently asked questions
We have been posting for six months with no result. Where do we start?
Not with the content, which is fourth by likelihood and the most expensive to test. Start by checking that nothing is genuinely happening: for two weeks, ask everybody who contacts you how they heard of you. A non-trivial share of cases resolves there — something was happening and was not being counted.
How do we know whether we are on the right platform?
Ask ten existing customers where they spend their time and what they use to find a supplier. Ten answers are enough, and they often contradict intuition, especially when you sell to businesses. The original choice was almost always made on what the owner uses, which is unrelated to where the customers are.
Our posts get reactions but no enquiries. Why?
Because there is nothing to do after seeing. One of three things is almost always missing: a reason to write now, an order of magnitude for the price, or simply a mention of what you sell. Add to one post in three a sentence saying what to do and how to reach you — not to all of them, which would make the page unreadable.
How long before concluding?
Below one post a week sustained for several months there is no conclusion to draw: the absence of a result is the expected result. Each cause test then runs six weeks, with one change and nothing else modified, and two figures recorded before and after — conversations started, and enquiries mentioning having seen you.
Should we change supplier?
Rarely first, because it is not a diagnosis. If the cause is reply time or a missing action, it survives the change intact and you will conclude twelve months later that the channel does not work. Do the diagnosis first; it will say whether the problem is at the supplier’s end or yours.
Can we decide to stop?
Yes, and it is sometimes the right decision: for some trades customers do not search that way, whatever the quality of execution. But do not leave the accounts as they are — an account abandoned eight months ago reads as a business that has closed. Update the information, publish one last permanent thing, and leave it as a business card.
Where we come in
The median delay before your first reply is a hard figure and it stings the first time. What it does not say is where to start.
- We run the diagnostic in the written order, skipping nothing even when it is slow.
- We change one thing between two measurements, and date which one.
- We tell you to stop the channel when the diagnostic says so.
For some trades this channel simply does not work, and nobody should tell you otherwise: six months of posting proves it far too late.
Read next
Social media management: you are buying a cadence, not a result
It is the only service in the pillar billed by volume of output. Twelve posts a month is not a result, it is a quantity.Your side of it: what has to be true at your end for this level to produce anything
At this level the outcome is decided at your end more than at ours. Advertising is half an act, and you hold the other half.Content marketing: the one asset you keep
Everything else in this pillar is rented and stops with the invoice. A written page stays, provided it answers a question somebody genuinely asks.
Let us talk about your project
A free audit, no commitment: we look at your online presence and tell you what is holding it back.