When we mapped what was actually being said about the brand online, the shape of it was unflattering and useful. Mentions spiked on financial results and trade shows. Underneath the spikes, the standing conversation was sales offers and service complaints. There was very little of anything else, and almost nothing happening on channels the brand did not own.
That is a common position for a large appliance manufacturer and it is frequently misread. The instinct is to treat it as a content problem — the brand is not saying enough interesting things — and to commission more content. The diagnosis I would defend is different. When your feed is a queue of people with broken machines, content is not the constraint. The queue is. Nothing you publish will be read in a context you have not first cleaned up.
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.
Running the complaints as an operation
So the first discipline was operational rather than creative: complaints received and complaints closed, counted every month, tracked as a pair.
Counting them as a pair is the whole point. Received on its own is a measure of how much noise you are under, which tempts everyone into trying to reduce it. Closed against received is a measure of whether the operation is keeping up, and it is the only version of the number that tells you anything you can act on. A rising complaint count with a rising closure rate is a healthy month. A flat complaint count with a falling closure rate is a bad one, and it looks identical in a report that shows volume alone.
Once that ran monthly, the reputation work stopped being an argument about tone and became a question of throughput.
Where the brand actually stood
Alongside it we tracked share of voice by product category against the competition, which produced a more honest picture than a single brand-health number.
In refrigerators the category conversation was dominated by a competitor whose share came substantially from press and blog coverage. In washing machines the brand did better, ahead of another domestic manufacturer — but those conversations were themselves largely sales offers and service issues, so the better ranking described a bigger queue rather than more affection. Different competitors lived on different platforms: one used a single microblogging channel heavily, another used social for its televisions, another for engagement.
Reading share of voice by category and by platform rather than in aggregate is what stops a brand from copying a competitor's tactic into a channel where that competitor's audience is not.
The premium launch, in the same feed
Against that background came the launch of a premium refrigerator, with a straightforward instruction: generate leads, and move the brand's imagery upmarket.
The strategy treated the product as a want rather than a need. Utility came second; the appeal was the object itself. The mechanism was association — an application that let people explore each feature by pairing it with a comparable luxury from another category, the panel with high-end touchscreen hardware and so on, with the surrounding content borrowing from sports cars, private jets and luxury experiences. The creative was minimal and clean, aimed at a narrow, affluent audience through tightly targeted placements.
The advertising performed: better than one per cent click-through on comparison sites, and close to one and a half per cent on behaviourally targeted placements, which are strong numbers for banners. Sales were satisfactory rather than remarkable, and the communities showed a clear appetite for the aspirational material.
The tension is the interesting part, and it is the reason I would put the two halves of this account on one page rather than two. A brand asking to be seen as premium is doing so in the same feed where it is answering service complaints about mid-market machines. Both are true and neither can be hidden from the other. The only version that works is the one where the complaints operation is visibly excellent, because a company that fixes things quickly has earned the right to sell an expensive object.
What I would do differently
I would have tied the complaint data to the product roadmap and said so publicly.
We had, every month, a structured record of what was going wrong with which machines, and we used it to run a support operation. That same record is product intelligence, and a manufacturer that visibly changes what it builds because of what its service queue says has converted its biggest liability into its most credible marketing claim. We were measuring the queue in order to keep up with it. The higher use was to let it change the product.