How to actually measure what works on organic social
Most brands posting on organic social have the same problem. They can see that a post did well. They can't tell you why.
The engagement numbers are right there in the dashboard. Likes, saves, reach, comments. But ask a sharper question, like whether your education videos beat your product videos, or whether talking to camera gets more saves than a product-only shot, and the dashboard has no answer. Not because the answer isn't in there. Because nothing is labelled, so nothing can be compared.
We built a fix for Balmonds recently: a tagging framework for their organic social. This post is how we did it, and why it's worth doing for any brand that's posting a lot and learning little.
Why the raw numbers don't tell you much
Organic reach is hard, and getting harder. Instagram engagement fell about 17% year on year (YoY) into 2026, with a median engagement rate near 1.12%, and it varies a lot by format. So you can't afford to guess. Every post is a small bet, and you want to know which bets pay off.
The trouble is that a standard social dashboard reports by post and by date. It tells you the Tuesday Reel got more saves than the Thursday carousel. It doesn't tell you what those posts had in common with your other winners, because the posts aren't grouped by anything that matters. Topic, audience, format, who's on screen, whether someone is talking to camera. None of it is recorded in a way the tool can read back.
So you end up with a feed full of numbers and no pattern. You repeat what felt good, not what worked.
The thing most brands miss
When brands do group their posts, they group them by topic. Education versus product. That's useful, but it's only half the picture.
The bigger driver is often how the post was made, not what it was about. The setting. The subject. Whether there's narration or text on screen, trending audio or original sound, a real person or just the product. In paid advertising this is well established: Nielsen puts creative at around half of the sales lift from advertising, the single biggest factor. The same logic applies to organic. The production style of a post often decides whether it lands.
Most brands never measure it, because they have no way to. That's the gap the framework closes.
What we built for Balmonds
Balmonds use Sprout Social to manage their organic channels. Sprout can tag posts, but a tagging tool only helps if the tags are consistent and mean something. Theirs had drifted: 29 loose tags sitting unclassified, a mix of shorthand product codes and one-off labels, with no structure holding them together.
So we built a proper framework. It works in two layers.
The first layer is the strategic one: six groups that capture what a post is and who it's for. Content pillar (education, product, routine, brand story, customer content, community, press). Product featured, so you can isolate performance by range. Audience or use case. Format. Funnel objective, from awareness through to conversion. And source, where the asset came from.
The second layer is the one most brands skip: production style. Four more groups covering the visual style, who or what is in the frame, who made the asset, and the on-screen elements. That last group is multi-select, so a single post can be marked as having narration, captions, a talking head, and a call to action (CTA) all at once.
Every post gets one tag from each strategic group and as many production tags as apply. We kept campaigns in Sprout's own built-in Campaigns module rather than rebuilding it as custom tags, so the brand still gets Sprout's campaign-level reporting. We set a strict naming convention so the library stays clean, and we mapped the 29 old tags into the new structure instead of throwing them away.
It is not complicated. It is just deliberate. The whole point is that every post is recorded the same way, every time, so the data can be read back later.
The actual tags, and the naming convention
Here's the framework we built, in full, so you can see how it's put together.
Every tag follows the same format: Category_Value. An underscore between the two, PascalCase for the value (each word capitalised, no spaces), and no hyphens. So Pillar_Education, not "pillar education" or "pillar-education". Sprout's tag search is case-sensitive in reports, so you set every tag up in the tag manager before the first post goes out, and you turn off ad-hoc tag creation so nobody invents a near-duplicate later.
Two of the codes use shorthand. UGC means user-generated content (customer photos, reviews, and reposts), and AIAssisted means generative artificial intelligence (AI) was used to make the asset.
The strategic layer, one tag from each:
1. Content pillar, the theme of the post: Pillar_Education, Pillar_Product, Pillar_RoutineHowTo, Pillar_BrandStory, Pillar_UGC, Pillar_Community, Pillar_Press.
2. Product featured, so you can isolate performance by range: Product_Hero, Product_Moisturiser, Product_Cleanser, Product_Serum, Product_Body, Product_Multiple, Product_None.
3. Audience or use case: Audience_Eczema, Audience_BabyChild, Audience_DryItchySkin, Audience_Psoriasis, Audience_Sensitive, Audience_General.
4. Format: Format_Reel, Format_Carousel, Format_Static, Format_Story, Format_Video, Format_UGCRepost.
5. Funnel objective: Funnel_Awareness, Funnel_Consideration, Funnel_Conversion, Funnel_Retention.
6. Source or origin: Source_Organic, Source_UGCCreator, Source_Influencer, Source_Press, Source_Repurposed.
The production layer, one tag from collections 7 to 9 and as many as apply from 10:
7. Visual style: Style_Lifestyle, Style_Studio, Style_FlatLay, Style_InTheWild, Style_Macro, Style_BehindTheScenes, Style_Graphic.
8. Subject, who or what is in frame: Subject_Model, Subject_Baby, Subject_Hands, Subject_Skin, Subject_ProductOnly, Subject_TeamMember, Subject_Pet, Subject_Multiple.
9. Production source, who made it: Production_InHouse, Production_Agency, Production_UGC, Production_Creator, Production_Stock, Production_AIAssisted.
10. On-screen elements, multi-select, apply all that are present: Element_TextOverlay, Element_Captions, Element_Narration, Element_TalkingHead, Element_TrendingAudio, Element_OriginalAudio, Element_NoSound, Element_PriceOrOffer, Element_CTAOnScreen, Element_BeforeAfter.
Campaigns sit in Sprout's own Campaigns module rather than in a custom collection, so tags like EczemaWeek, BFCM2026 (Black Friday and Cyber Monday), a Launch_[Product] tag per launch, BCorp, and seasonal pushes like Seasonal_Winter get the benefit of Sprout's campaign-level reporting.
The product tags above are generic examples, and the audience tags are illustrative. Those two collections are the ones you tailor: your own ranges, your own customers. Everything else, the pillars, formats, funnel, styles, subjects, and on-screen elements, transfers to almost any ecommerce brand more or less as is.
The questions it lets you answer
Tagging is boring on day one. The payoff comes after about 60 to 90 days of consistent posting, when there's enough labelled data for Sprout's tag reports to compare like with like. At that point you can ask real questions and get real answers:
Do education Reels outperform product Reels?
Does talking to camera drive more saves than a product-only post?
Do rough, in-house shots beat polished studio shots for product?
Does narration beat text-on-image for educational content?
Is trending audio worth chasing, or does original audio hold attention better?
Which content pillar earns the most reach and saves?
Every one of those is a decision you make most weeks anyway. The difference is you'd be making it from evidence, not instinct. And because the framework tags production style, the winners feed straight into paid creative testing. Your organic feed becomes a low-cost testing ground for the ads.
Why this is worth doing
If you're posting a lot and struggling to improve, this is usually the missing piece. A few reasons it works.
You stop guessing. The framework turns a busy content calendar into a structured test, so each month teaches you something you can act on.
You separate topic from how it's made. Knowing that talking-head education posts beat product flat-lays is worth more than knowing education does well in general, because it tells you what to brief next.
It builds over time. The longer you run it, the more the data is worth. Early on you learn the obvious things. Later you spot the patterns competitors never bother to look for.
It costs almost nothing. The tool you already pay for can do this. Most brands just never set it up properly.
How to set it up
If you want to do this yourself, the sequence is straightforward.
Build the collections in your tool's tag manager before the next post goes out.
Add every tag, spelled exactly, and turn off ad-hoc tag creation so nobody invents new ones on the fly.
Triage any existing tags into the new structure, and archive the dead ones.
Keep campaigns in your tool's native campaigns feature, not as custom tags.
Brief whoever publishes, so every post is tagged when it goes out, not weeks later from memory.
Then keep it honest. Tag at the point of publishing. Review the library monthly for the first quarter, then quarterly. A 15-minute check every couple of weeks keeps standards high while the habit beds in. The framework is only as good as the consistency behind it.
How Sector 106 does it
Building reporting frameworks like this is part of our Strategy & Operations work. For Balmonds we audited the existing Sprout set-up, designed the ten collections across the strategic and production layers, wrote the naming conventions, mapped the old tags, and set the review cadence. The deliverable is a short reference guide their team can publish against, with no further input from us needed.
We do the same across channels: structure the data first, so the analysis afterwards is actually possible. It's quiet, behind-the-scenes work. It's also the difference between a brand that learns from its content and one that just produces more of it.
The honest test
Two questions:
Can your social tool tell you, with numbers, whether talking-head posts beat product-only posts?
Could you say which production style drives the most saves for your brand?
If the answer is no, the data isn't broken. It just isn't labelled. That's about a week of set-up, and then it pays you back every month after.
Want your organic social set up to actually teach you something?
We build tagging and reporting frameworks for ecommerce brands, in Sprout Social or whatever tool you use. Senior people, 28 years in.
Or read more: Strategy & Operations →
