Social Ads Strategy from (un)Common Logic

Social ads still move markets, but only when you respect the math, the message, and the moment a customer is in. I have watched brands double revenue on the same media budget with smarter creative, tighter objectives, and a habit of ruthless testing. I have also seen great products stall because the team tried to win with a single lookalike audience and a dashboard of vanity metrics. A strong social strategy is not one trick. It is a set of operating principles, stitched into daily execution, and measured with a clear-eyed view of causality.

This article lays out how we at (un)Common Logic approach social advertising across Meta, TikTok, LinkedIn, YouTube, Pinterest, and Reddit, with an emphasis on e‑commerce and B2B lead generation. The goal is a program that can scale spend 2 to 5 times while holding, or improving, blended return.
Start with the economics, not the ad unit
Your revenue formula in paid social reduces to three levers: traffic volume, conversion rate, and value per order or lead. The platform matters less than how precisely you choose a conversion objective, align creative to the moment in the funnel, and fund the system with enough signal density for learning. If those are set, CPMs and CPCs become inputs you manage. If they are not, no audience hack will save you.

We frame the economics at two levels. First, channel level targets that define guardrails, such as a blended MER between 2.5 and 4.0 for a growth stage DTC brand, or a cost per qualified lead below one third of average contract value for B2B. Second, campaign level targets that reflect intent: prospecting can carry higher CPAs if remarketing and email recapture pull the blended cost back inline. That view encourages the right tradeoffs. You will fund prospecting into rising CPCs if it feeds high LTV cohorts, and you will pause remarketing if you find it simply harvests conversions that would have happened from owned touchpoints.
Objectives are steering wheels, not labels
The objective you select tells the platform what to hunt. A conversions objective with a purchase event gives the algorithm a clean target, but only if you have signal density. As a rule of thumb, you want 50 to 100 tracked conversion events per ad set per week for stable learning. If you are below that, consider consolidating ad sets, optimizing for an upper funnel event temporarily, or using Advantage+ Shopping or Sales Campaigns on Meta that pool signals across creative. With B2B, optimize to pipeline, not raw leads, once you have enough data. Start with a lead objective if volume is scarce, then shift to qualified lead or opportunity as the platform learns.

A common mistake is to split budget across too many small ad sets and objectives, starving the system of data. Another is to run traffic or engagement campaigns for months because the reported CPC looks attractive, even though downstream revenue never materializes. I would rather pay a 40 percent higher CPC for clicks that convert at double the rate than scale cheap traffic that bounces.
The funnel is real, but it is not a staircase
Many decks present the funnel as three neat boxes: awareness, consideration, conversion. Real customers slosh back and forth. A TikTok that makes someone laugh can spark a Google search two days later, then a Pinterest browse, then a purchase through a Meta dynamic ad. That path scrambles last click reporting. You need a mental model that treats social ads as both demand creation and demand capture.

Practically, we design creative and offers for six customer moments: unprompted discovery, affinitized curiosity, active comparison, proof seeking, cart friction, and post purchase advocacy. The same platform can serve any of those moments if the creative and call to action fit the intent. For example, a product teardown video can drive first touch interest at the same time a testimonial carousel handles skeptics who already visited the site.
Creative is a performance variable, not a costume
On social, creative is the main performance driver once your targeting and spend thresholds are competent. The difference between a top decile and median ad often shows up as a 20 to 60 percent shift in click through rate and a meaningful lift in conversion due to higher intent traffic.

We break creative into jobs, not formats. A hook should stop the thumb within the first second. The body should resolve the tension you created with the hook. The proof point should anchor belief, using numbers, social proof, or a visual demonstration. The close should make the next step feel small. Within that structure, we rotate variations that test angle, not just color or button text. Here are four angles that regularly produce winners:
Friction removal: show a pain being solved in under eight seconds, with concrete before and after. Social currency: make the buyer look smart, resourceful, or in the know. Cost logic: compare price to an obvious alternative or quantify the waste of inaction. Authority transfer: borrow belief from customers, experts, or third party awards, with specific claims.
For short form video, production value matters less than clarity and energy. A well lit smartphone clip with crisp audio often beats a polished brand film. Subtitles should be burned in, and any on screen text should be legible in the first frame. Static ads still work for catalog or offer communication, but they rarely break into new audiences. If you must choose where to invest, build a repeatable UGC and founder faced pipeline that delivers two to three new videos per week per core product line.
Targeting in a privacy first world
Broad targeting has become the default on Meta and TikTok because the algorithms are good at finding converters once they have signal. That said, broad is not a free lunch. It works when your pixel, conversion schema, and creative quality create strong training data. If your funnel is leaky or your site is slow, broad amplifies the problem.

Lookalikes still help in B2B or niche markets where the base rate of likely buyers is low. We have seen 1 and 2 percent lookalikes from closed won deals on LinkedIn and Meta outperform broad by 20 to 40 percent in cost per opportunity when deal cycles are long. Interest targeting can be useful in early testing to identify angle resonance, but we avoid slicing interests so thin that each ad set drops below learning thresholds.

Retargeting has changed since iOS 14.5. Window lengths are shorter, match rates are lower, and overlap with owned channels is higher. We run retargeting with tighter burn windows and creative that assumes knowledge. Dynamic product ads remain efficient for e‑commerce if your feed quality is high. For lead gen, we bias toward retargeting with proofs and objection handling rather than repeating the same top of funnel spot.
Platform choices are business choices
Meta remains the workhorse for most DTC brands because of scale and mature optimization. TikTok brings incremental reach and often cheaper CPMs with more variable conversion rates. Pinterest, Snap, and Reddit can be productive for certain categories where the native behavior matches your product use cases. LinkedIn is expensive on a CPC basis, but for high ACV B2B it can pay for itself with tight account lists and content that addresses evaluation criteria.

YouTube works best when you think of it as upper funnel TV with measurable intent spillover, not a direct response clone of search. The winning play is to run non skippable or skippable in stream to seed demand, then measure both branded search lift and assisted conversions in blended reporting. For brands with product demonstrations that benefit from longer narratives, YouTube can pull double duty as awareness and proof.
Budgets that respect learning and scale
We size budgets from target outcomes backward. If the revenue plan requires an incremental 2 million dollars this quarter and you hold a blended MER of 3, then your allowable media spend is about 666,000 dollars. From there, allocate a percentage to social based on historical contribution to last click and assisted revenue. Within social, prospecting should generally carry 60 to 80 percent of spend for growing brands, with retargeting and loyalty rounding out the rest. For B2B, allocate at least half of spend to content or event promotion that precedes the lead form. The longer the sales cycle, the more you must fund education.

Avoid whipsawing budgets. Algorithms do not respond well to daily 30 percent cuts or boosts. We make changes in 10 to 20 percent increments and let new levels stabilize for three to five days unless there is a clear tracking break or offer change. During promotions, pre‑fund prospecting one to two weeks ahead so more users enter retargeting windows before the sale. Then, on the promotion days, raise retargeting frequency caps slightly and refresh creative to avoid fatigue.
Measurement that keeps you honest
Attribution on social will always be messy. Treat the in‑platform numbers as directional, not gospel. For performance reviews and planning, rely on a blended view: channel level spend against total site revenue and lead pipeline, adjusted for seasonality and promotions. Use post purchase surveys to triangulate, and run geo experiments or holdouts when budgets permit. Even a small test where one region pauses prospecting for two weeks can reveal incrementality that pixels cannot.

On Meta, you will often see conversion modeled events fill in over the following days. Do not yank budgets based on day one ROAS if your historical lag is three to seven days. On TikTok, expect more view through assists relative to click through conversions. Judge creative on both in‑platform performance and how it changes branded search volume or direct site traffic.

For B2B, align CRM stages and UTM hygiene before you scale. A lead with a work email is not a success if it never hits qualified stages. Pipe opportunity and revenue data back to the platform with offline conversions, then optimize toward the highest stage you can feed with at least 50 events a week. When that is not feasible, build a lightweight model that predicts lead quality from form fields and early behaviors, then use that as a proxy.
Offers and on‑site experience carry more weight than most teams admit
A 10 percent discount rarely moves a high consideration purchase. A value add, such as a free accessory or extended trial, can supply a stronger nudge without training customers to wait for coupons. For subscriptions, stack a low friction trial with an early reorder reminder sequence that highlights the first month’s outcome. On B2B forms, cut fields ruthlessly. Every extra question raises CPA. If sales needs qualification, use progressive profiling or route to a research SDR who can enrich firmographic data after the form is submitted.

Site speed will make or break paid social. An otherwise great campaign can lose half its power if the mobile site takes five seconds to load. Aim for the first contentful paint under two seconds on 4G and tighten third party scripts. Streamline checkout and remove surprise shipping costs early. We have seen conversion rate jump 20 to 40 percent from basic UX fixes, which then makes every ad dollar more potent.
Creative production cadence
High performing social programs act like newsrooms. They ship new creative weekly, driven by a brief that ties audience insight to an angle and a format. They mine comments for objections. They turn customer support transcripts into scripts. They calendar product launches, seasons, and cultural moments, but they also leave room for opportunistic content when a hook emerges organically.

A workable cadence looks like this: ideate on Monday with three to five angles mapped to funnel moments, script and shoot Tuesday to Thursday, edit on Friday, and launch on Monday morning with prewritten variants for the first round of testing. Keep raw footage organized so you can resurface winning scenes months later with fresh hooks.
Testing that respects statistics and attention
Teams frequently burn budget on tests that cannot produce confident reads. Better to test fewer things well. When running A/B creative tests, isolate one dimension when possible, such as hook or offer, and feed both variants similar audiences. Let the test run through at least one learning phase reset. Resist calling winners on day one.

Here is a simple cadence that has served us across accounts:
Pick one primary variable per week, such as hook angle or offer framing, and one secondary variable, such as headline line break or end card CTA. Launch two to three primary variants and two secondary tweaks for the strongest candidate, for a total of three to five active ads. Set minimum spend thresholds that ensure each variant can reach 1,500 to 3,000 impressions before you judge click through rate, and enough clicks to judge conversion rate. Archive clear losers quickly to concentrate spend, but let near ties run an extra day to break variance. Roll the winner into your evergreen pool and immediately brief the next round that builds on what you learned. A quick diagnostic checklist when results stall Does the platform see enough conversion events per ad set per week to learn, or are we starving it? Are we using a conversion objective aligned to the economic target, or did we slip into traffic and engagement comfort? Has creative fatigue set in, with frequency above 2.5 in a week and falling click through rate? Do landing pages and offers match the promise of the ad, with reasonable speed on mobile? Are budgets swinging too wildly for the algorithm to adapt, or are promotions confusing the baseline? B2B specifics that change the playbook
With B2B, the hardest part is not getting a lead. It is getting a decision maker back to the site after the first click. Content that hits evaluation criteria wins: security one pagers, ROI calculators, integration demos, and short testimonials from relevant industries. LinkedIn’s lead gen forms can work, but do not rely on them exclusively. Route form fills to a sequence that includes calendar booking and a light content drip. If your ACV justifies it, sync a target account list and layer in company size, industry, and seniority. Guard against vanity metrics by reporting on opportunity stage and projected value, not top of funnel CPL.

We have seen strong results from promoting a three part webinar series that unpacks a complex migration, then retargeting attendees with short clips addressing common objections. The cost per opportunity can drop 20 to 30 percent versus cold calls alone because your audience shows up warmed and educated.
E‑commerce moves that stack advantage
On the DTC side, product and price carry the day, but you can widen the edge with smart programs. Feed quality for dynamic ads matters more than most give it credit for. Fix titles, add attributes, and include real promotional pricing rather than lazy strike throughs. Use Advantage+ Shopping or Performance Max equivalents with guardrails while running controlled prospecting campaigns that test angles. When you find a winner, back it with more spend and a refreshed cut every week to keep frequency in check.

Bundles and limited runs create urgency. If inventory is tight, communicate count thresholds honestly. Early access for SMS or email subscribers builds your owned list, which then softens MER pressure after campaigns end. For repeat purchase categories, use ads to reset the habit window. A well timed reorder prompt featuring a testimonial about long term results nudges retention better than a generic 10 percent off.
Team, process, and culture
Results come from a team that can move fast without losing the plot. Media buyers, creative strategists, editors, and analysts should sit close enough that feedback loops are measured in hours, not weeks. Daily standups keep campaigns honest. A shared dashboard that updates spend, ROAS, CPA, and qualitative notes prevents siloed decision making. Most importantly, give someone the authority to pause a beloved ad if it stops working. Attachment kills performance.

We also document decisions. If we lower bids, we note the reason and the expected effect. If an angle wins, we capture why we think it resonated and what to try next. Over time, this builds a proprietary library of what your market responds to, which becomes a real moat.
Tooling without tool worship
The platform tools, plus a clean analytics setup, will cover https://milonnrk144.tearosediner.net/the-un-common-logic-data-governance-playbook https://milonnrk144.tearosediner.net/the-un-common-logic-data-governance-playbook most needs. Useful additions include a creative analytics layer that tags hooks and angles, a UTM builder that enforces standards, and a database or spreadsheet that tracks experiments. For B2B, CRM integrations that pass offline conversions back to ad platforms are worth the effort once volume justifies it. Resist the urge to chase every shiny SaaS that promises perfect attribution. None exists. Spend that energy on clean data and faster creative cycles.
Case sketches from the field
A DTC skincare brand at 5 million dollars annual revenue grew paid social from 120,000 to 300,000 dollars a month while holding a blended MER of 3.2. The unlocks were a tightened creative brief that shifted from glossy lifestyle to pore level demonstrations, a faster site on mobile, and a budget structure that weighted prospecting at 70 percent with a rotating library of hooks focused on problem solving. Single video CTR lifted from 0.8 to 1.6 percent, and the site conversion rate rose from 2.0 to 2.7 percent, which compounded into scale.

A B2B SaaS company selling compliance software struggled with lead quality at a 160 dollar CPL on LinkedIn. We rebuilt the flow to promote a three step buyer’s guide with a clear checklist of regulatory gaps. The lead form dropped to 110 dollars, but the bigger gain came from routing those leads into a webinar series with a security architect. Opportunity rate rose from 7 to 14 percent, and cost per opportunity fell by 40 percent. Feeding closed won data back to the platform let us move the optimization target to opportunities, which stabilized performance.

A home goods retailer relied heavily on retargeting and saw declining returns as match rates worsened. We reallocated budget to upper funnel creative featuring side by side comparisons of build quality, then paired that with a price match guarantee on the product page. Prospecting CPA rose by 18 percent, but overall revenue grew 35 percent on essentially flat spend because new customer volume increased and email captured more of the follow up.
Tradeoffs and edge cases
Every rule will break somewhere. Broad targeting can backfire for hyper niche products with tiny addressable markets. In that case, lookalikes from high quality seeds and tight interest stacks can beat broad until you accumulate enough data. Short promotion windows can justify aggressive daily budget swings, even if that irritates the algorithm, because the human behavior response outweighs the machine’s discomfort. Likewise, if your product depends on seasonal context, like tax filing or back to school, your testing cadences must compress to capture the moment.

Beware of false positives during sales. Retargeting will look heroic when you drop prices across the site. Once the sale ends, those same ad sets often fall below your thresholds. We make a point to analyze promotion weeks separately and measure their halo in the following weeks before changing baselines.

Finally, know when to pause. If your tracking is broken and the platform is flying blind, you are feeding a machine with the wrong target. Fix the pipe first. When creative is exhausted and performance decays, step back for a week to rebuild your angle library rather than try the tenth tweak to a tired spot.
What makes a social program resilient
Resilience comes from four habits. First, you align objectives with economic targets and adjust them as signal density grows. Second, you treat creative as a system, not a project, with steady inflow and ruthless pruning. Third, you measure with humility, using blended revenue and targeted experiments to validate lift. Fourth, you train the whole team to spot weak signals early: a slowing hook rate, a rising checkout drop off, a shift in comment sentiment.

At (un)Common Logic, that is the spine of our social ads strategy. The platforms will change. Targeting knobs will come and go. But teams that understand the economics, respect the algorithm without worshiping it, and publish work that earns attention will keep finding their customers at scale.