At PingPong, the cross-border payments company where I was Marketing Manager, Ads and Lifecycle from 2021 to 2023, I was also the social program: one person writing, designing, scheduling, and reporting, with no dedicated designer and no media budget worth bragging about. My first attempt looked the way most small-team social looks: post when something's ready, boost whatever seemed strong, watch the reach number climb, repeat. It produced a busy feed and nothing I could point to in a review. Nobody told me headcount was off the table. I decided the system had to work before I asked for anyone, so the campaign had to run as a system, because the system is the only part of the operation that shows up every week whether or not you're having a good one.
Message before visuals
I write the message down before a single asset exists. I keep the document as an internal library: the key messages I'm committing us to, the misconceptions the audience holds, the language patterns we can repeat without wearing them out, and the narrative angle that ties it together.
This looks like overhead until week three. Then it's the reason a caption, a carousel, a short-form video, and a repurposed cut all sound like the same campaign. Write the message once and everything downstream gets cheap.
Templates that cut production cost
One person can only sustain a campaign where each asset is cheap to make. So I built the frames next: headline blocks, quote card designs, explainer frames, carousel skeletons, a few video layouts. Built once, filled many times.
I built the frames to get five versions out without waiting two weeks on a designer I did not have. Once the frames exist, making a post means filling one in, design stops blocking every post, and the only thing that still has to be good each week is the message going into the frame.
One idea adapted to four platforms
Cross-posting the same asset everywhere reads as laziness to the audience and performs like it. Here's how I adapted one idea from the payments campaign, the social-proof claim that our platform was trusted by 10,000 businesses:
- LinkedIn, a short narrative post: what those businesses move across borders and why they switched, with the number as the closer.
- Instagram, a quote card straight from the template library: the customer count as the headline, one line of context, nothing else.
- Facebook, problem first: the cost and delay of legacy cross-border transfers, then the proof that 10,000 businesses had already moved on.
- Short-form video, the number in the first second as the hook, then one concrete customer scenario, under thirty seconds, nothing spare.
I wrote that claim once, in the library, and the templates made the four shapes cheap.
Distribution is a rhythm
Posting “when something's ready” produces scatter. I know, because that was my first system, if it deserves the word. What replaced it is a weekly sequence: warm-up content to open, then one educational post, one credibility post, one practical post, something lightweight, and then something deeper. Stacked in that order, a week of posts builds an argument. One person schedules all of it, so the rhythm doubles as the production plan. I always know what kind of post I owe next, which is half the battle when nobody is assigning you anything. Once the frames were in place that held at four or five posts a week across the platforms, for about three months.
Testing ad angles without a testing budget
The paid side of a resource-constrained campaign lives or dies on angle testing. An angle is the core premise of the ad, the problem it names and the emotional trigger it pulls, distinct from the image, the headline, or the CTA. My first tests changed all of those at once, which is a reliable way to spend money and learn nothing. So I split it into two phases.
Phase 1: find the image
Run 10–20 images on click-through rate at a small budget, changing nothing but the image. One ad set, duplicated 9–19 times. This tells me which image people stop on, before copy muddies it.
Phase 2: find the angle
Take the winning image, duplicate the ad set, and vary one angle per ad, identical everywhere except the copy. Three days, roughly $20 per ad set carved out of the paid budget already running, since social had no line of its own, then scale or cut on the early indicators. In one round for a payments product, the three candidate angles were:
- Discount, “Save 20% on Cross-Border Payments”
- Benefit, “Simplify Your Payment Process in Minutes”
- Social proof, “Join 10,000 Businesses Trusting Our Payment Solutions”
Social proof beat the next-best angle on click-through rate by 15% relative. The test measured clicks and nothing past them. Each set drew a few thousand impressions and a few dozen clicks, so the read was directional. I pushed the winner into regular rotation without a second round, and nothing downstream of the click was wired back. I wouldn't have guessed it, and that's the argument for a $20 three-day test: cheap enough to run again before I act on it.
Close the loop: social is an input too
A post's life doesn't end at publish. I fed posts into newsletters, video scripts and sales decks, and mined strong comments for new topics and customer quotes. Sales and support hear the questions I never see, so I asked them for a list every Friday. Because the message stayed the same across social, email and live conversations, each channel reinforced the others.
Metrics that mean something
I stopped reporting reach and impressions. I watch saves, replays, comments with real questions, longer dwell time, and whether people share the piece privately. Each of those tells me to do something, or I stop watching it. When people saved a post and shared it privately, I expanded it, rewrote it longer, or cut it into a video using the same template frames. A comment with a real question became the next week's educational post. A post with big reach and nothing else, I retired, however good it looked on the platform dashboard. I never kept absolute values on any of them. I read them as a trend.
The honest limitation: the system makes assets cheap to produce and tests cheap to settle, but it doesn't make attribution clean. I can defend the 15% angle win and the three $20 ad sets behind it, though $60 only buys a direction; I can't hand you a tidy revenue line for organic social, and I won't pretend to. What I can show is where every dollar and hour went, and what decision each one bought.
Where to put the first $20
If you run social by yourself, write the message library before you make anything, and build the templates before you post anything. Then put your first $20 into an angle test and leave the boost button alone.