The brief was recognisable. Conversation about the brand online was queries, complaint registration, negative comments and a steady supply of sarcasm, and the task was to move it toward something positive and experiential.
I led the team on this account and the pitch that won it. I was close to the planning throughout, and the execution was done by people I had hired.
Before proposing anything, we sampled what people were actually complaining about. That sample is the reason this study is worth reading, because it contradicted the brief.
Where the anger was actually coming from
Roughly three-quarters of the complaints — 74 per cent — were about after-sales service. Poor delivery accounted for about 14 per cent, and faulty products for about 12 per cent.
That distribution decides everything downstream. If most complaints had been about products, the answer would be merchandising and expectation-setting, both of which communications can influence. If they had been about delivery, the answer would be logistics and better tracking. But when three in four complaints are about what happens after something has gone wrong — the resolution process itself — the problem is that the company is bad at repair, and no amount of positive content sits comfortably on top of that.
The scale mattered too. On an average day more than thirty dissatisfied customers were complaining publicly, around six times the volume a leading technology brand was absorbing. Sentiment split hard by platform: on the company's own page around eighty per cent of comments were negative against twenty positive ones, while on the faster public network around thirty per cent were negative, with roughly 3,200 people saying positive things over the same window and about 1,600 conversations running neutral to negative across sixty days. Forums and video ran neutral. News ran mostly positive.
Reading that as one "sentiment score" would have produced a number that was true and useless. The useful reading is that the owned channel had become the complaints desk, which is exactly where the company had least room to manoeuvre and most obligation to respond.
Why the timing of the conversation matters in retail
Two findings from the shopping research shaped the recommendation.
The first: most buyers use social platforms before purchase, to get ideas and to compare, and a large majority settle on their preferred shopping destination immediately after their first encounter with a brand online. The decision is made early, on thin evidence, and it is sticky.
The second: shopping does not end at checkout. A meaningful share of buyers left a comment about the service they had used afterwards. So the post-purchase experience is not the end of a transaction — it is the raw material of the next person's first encounter.
Put those together and the case makes itself. In a marketplace business, after-sales service is the acquisition channel. It produces the evidence a prospective buyer meets at the exact moment they choose a destination, and it does so at a volume no campaign can outspend.
Segmenting buyers by what they can be asked to do
The segmentation split buyers by motive rather than demographics. Those who need to be seen discovering things first. Those who plan and buy deliberately and judge others by what they own. Those who buy sporadically and mostly on a friend's recommendation. And those who are loyal to a discount rather than to a seller.
The value of that cut is that each segment can be asked for something different. The discoverers will advocate publicly if there is something to be first to. The planners will produce considered reviews. The followers will convert on a friend's word but will not generate one unsupported. The deal seekers will do neither reliably, and building a programme that depends on their affection is a mistake.
What I would do differently
I would have made the complaint distribution the pitch.
We presented it as evidence for a communications strategy, when the finding was larger than the brief we had been given: the reputation problem was an operations problem, and the number proved it. The braver and more useful move would have been to hand the client a service diagnosis, offer to measure it monthly, and say plainly that the communications programme would be worth buying once the resolution process was fixed. A finding that inconvenient is usually the most valuable thing in the room, and it is the part everyone is tempted to soften.