Over a decade ago, I was putting my daughter to bed. One hand in the cot, tapping her to sleep. The other hand under the bed, phone face-up, finishing a caption I’d drafted earlier that day.
And I remember thinking: there has to be a better way than this.
The whole thing. My business success, the clients I was responsible for, the income that paid for everything — all of it hinged on whether this post would land. Whether my guess about what to write would be a good one. Whether the algorithm would decide to show it to people or not.
That didn’t seem fair. It didn’t seem rational. And it definitely wasn’t a systematic way to grow a business.
That was the moment Clue Labs started.
The Question Nobody Could Answer
I’d been building businesses on social media for years by then. I knew what I was doing — I sold two properties worth £2 million combined through Instagram DMs. I grew accounts that turned into real revenue. I had results I could point to.
What I couldn’t tell you was why it was working. I had patterns I’d noticed, instincts I’d developed. But when it came down to it I was making educated guesses, and my clients’ outcomes depended on how good those guesses were.
So I started looking at the tools. Not clients — tools. I went through most of what existed: schedulers, analytics platforms, listening tools, reporting dashboards. Some of them were genuinely good at what they did. Beautiful interfaces. Deep data.
But none of them could tell me what to post next. Not based on my specific account. Not based on what the algorithm was actually rewarding right now. They could show me what had already happened. They couldn’t tell me what to do about it.
They were measuring the past and calling it strategy.
The Incentives Problem
The more I looked at it, the more I realised this wasn’t an accident. It was structural.
Every major social media tool on the market makes money when you use it more. More posts scheduled. More reports generated. More features turned on. Their commercial model depends on your activity level.
Which means they are financially incentivised to tell you to do more. Post more. Test more. Spend more on ads when organic doesn’t work. That advice isn’t necessarily wrong. But it’s not neutral either. It’s shaped by what grows their business, not yours.
And then I looked at what the platforms had done.
Instagram, LinkedIn, Facebook, TikTok — they’ve all restructured around AI-driven discovery. Your followers don’t automatically see your posts anymore. The algorithm decides, based on predicted engagement, who deserves to see your content. Every platform became a discovery engine. Organic reach didn’t die — it became something you have to earn differently.
The existing tools hadn’t caught up. They could tell you your reach was down. They couldn’t tell you why, or what to do about it, in terms of the actual signals driving it.
Someone needed to build for the model as it actually exists. That was the gap.
What I Wanted to Build
Something that reads your data and tells you what to post next. Not a template. Not best practices. A recommendation built from your specific account, your specific audience, and what the algorithm is currently rewarding.
Something that works from the signals that actually matter — not reach and impressions, but engagement trend, save rate, consistency pattern, audience growth direction — and turns those into a clear content direction.
And something that, when it’s working properly, tells you to post less. Less of the right thing. Because the whole model being sold to brands — post constantly, always be testing, more content equals more reach — was never based on how the algorithm actually works. It was based on what the tools needed you to believe.
That’s what Clue Labs was built to be. The tool that’s on your side, not the platform’s.
What Building It Actually Looked Like
Longer than I planned. More expensive than I budgeted. One development team that wasn’t right before we found the right one.
There was a period where the product stagnated and customers churned. I’m not going to dress that up. It was hard — not as a metric, but because people had believed in the vision early and watched it not move fast enough. I understood why they left.
What kept me going was the work happening underneath. The architectural redesign. The AI pipeline. Getting the right technical leadership in. The work of actually mapping what a social media intelligence platform needs to look like at the level of the algorithm — not the dashboard, the actual model underneath it.
That work wasn’t visible to anyone. But it’s what made everything that came next possible.
Where We Are Now
Clue Labs v3 launches 1st May 2026.
Free tier: your Missed Opportunity Score and two post diagnoses a week. No credit card. No time limit. Real data, real insight, before you spend anything.
Paid platform: campaign recommendations built from your account data. Wild Card for one-off ideas when you don’t want a full campaign. Newsjacking for daily reactive briefs matched to your brand. Top Fans to see exactly who your most engaged audience members are. Clue Score as your single weekly measure of whether your account is building algorithmic momentum or losing it.
LinkedIn is in active development. More platforms to follow.
The category I’ve been building toward — Social Discovery Optimisation — is starting to be named in industry media. Marketers who’ve been watching their reach decline for two years are starting to ask the right questions.
I built Clue Labs to give them the answers. Because nobody should be finishing a caption one-handed in the dark, hoping their guess is good enough.
