Most performance marketing campaigns fail because of wrong audience, weak offer, poor creative, a landing page that doesn’t match the ad, tracking nobody trusts, or a follow-up process that lets a good lead go cold. Six weak links, and a campaign only needs to break at one of them to fail. That is the reason we don’t run campaigns platform-first. “We’re good at Google Ads” or “we’re good at Meta Ads” miss the business outcomes. At 8Spades, every paid campaign (performance marketing) we run, whether it’s for a real estate developer in Chennai, an industrial pump manufacturer selling to factories, or a D2C coffee brand, follows the same structured process. We call it the 6A Framework: Audience, Auction, Ad Creative, Action, Analytics, and AI. Six deliberate checkpoints, run in order, every time.
1. Audience: Who are you actually trying to reach?
Every campaign starts here, and it is the step most businesses rush through. “Women, 25-45, interested in fitness” is a demographic. It is not buyer persona. The real work is understanding how a specific person actually behaves before they ever see your ad. When we onboarded a leading global pump manufacturer’s industrial division in India, their existing lead generation was costing upwards of ₹10,000 per lead through their standard global playbook, a cost their sales team couldn’t justify. Rather than guess at targeting, we went back to their distributor network and asked pointed questions: What issues do clients actually raise around water treatment, temperature control, and industrial processes? Who are the real decision-makers, and at what stage do they get pulled into a sale? What kills a deal outright?
What came back reshaped the entire audience strategy. Returning customers searched for exact applications and pump models. New customers searched by problem statement phrases like “pump out chloroform from underground tanks” or “replace existing ordinary multistage pump with energy efficient.” Plumbing, HVAC, and MEP consultants turned out to be the hidden decision-makers on greenfield projects. None of that surfaces from a targeting dropdown menu. It surfaces from talking to the people closest to the actual sale. The distributors, sales teams, and existing customer conversations. For lead-generation businesses in particular (real estate, BFSI, B2B, healthcare, education), audience work also means accepting that the funnel doesn’t end at a form fill. A prospect who filled a form for a real estate project today might not be ready to visit the site for another three weeks. Audience strategy has to account for that gap, not just the first click.
Audience can be typically targeted through the following ways: keywords that they use to search/research a product/category/brand, demographics, geography, interests, online behavior, first-party data (customer data collected, stored, and managed directly), and lookalike audiences.
2. Auction: Where and how you actually compete for attention
Once you know who you’re targeting, the next question is where you show up, and on what terms. Search and social work fundamentally differently, and treating them the same is one of the most common mistakes we see businesses make when they bring performance marketing in-house. On Google, people are actively searching, so the job is to show up at the exact moment intent already exists. On Meta, people are browsing, not searching, so the ad’s job is to introduce a need they weren’t actively looking to solve.
For the pump manufacturer, that meant intent-based Search campaigns built around brand, product, and problem keywords, layered with interest-based Meta and LinkedIn targeting by profession and industry (Pharma, Auto, Chemical, F&B), each with its own landing page. The result: a cost per Marketing Qualified Lead roughly 75% lower than their existing global campaigns, simply by rethinking which platform should be doing which job, and bidding accordingly.
The auction step is also where volume and quality get weighed against each other, something we push every client to think through before a campaign launches, not after. We frame it plainly for teams: would you rather have 1,000 leads with 100 genuine ones, or 500 leads with the same 100 genuine ones? The second option is nearly always the better business outcome, even though it looks worse on a leads-generated report. Bidding strategy has to be built around that answer, not around vanity volume. Bidding strategies can be used to maximize impressions, reach, clicks, conversions, or ROAS based on the context of the campaign.
3. Ad Creative: Earning the click you’re bidding for
Winning the auction only gets you the impression. Someone still has to stop scrolling. This is where we push back hardest on a common instinct: businesses obsess over content polish far more than the data justifies. Good content with no distribution is invisible. Average content with strong distribution creates real impact. It’s the pairing that wins, not creative quality in isolation.
That said, creative still has to say something worth stopping for, and what “worth stopping for” means changes completely depending on the business. For a motorsport-driven brand, that meant leading with raw racing footage: real rally wins, real conditions, because the credibility was already built into the sport itself, and the audience could tell the difference between manufactured content and the real thing. For a commercial vehicle brand competing on durability and engineering, the winning creative leaned into demonstrable proof: trucks pushed through conditions Indian roads actually throw at them, not polished studio shots.
The discipline we apply across every account is matching creative format to funnel stage. Top-of-funnel creative on Meta needs product discovery, UGC, and influencer content doing the introducing. Bottom-of-funnel creative should already assume familiarity: promotional and sign-up focused. Running the same generic brand ad across every stage of the funnel is one of the fastest ways to waste a media budget that was otherwise well-targeted.
4. Action: What happens after the click
This is the step most businesses underinvest in, and it’s the one that decides whether the previous three steps actually convert into revenue. We call it “Action” because what counts as the meaningful next action is different for every business model, and treating them identically is a mistake we actively correct for in every new client conversation. We think about it across three distinct journeys:
- Lead-generation businesses (real estate, BFSI, B2B, education, healthcare): the chain is Ad → Form Fill → Sales Call → Sale. A form fill isn’t the conversion, it’s the first step of a process that often plays out over weeks. This is why we insist on connecting ad platforms to a client’s CRM before we scale spend. Without that connection, a business is optimizing for leads instead of genuine leads, and those are frequently not the same people.
- App user acquisition businesses (gaming, fintech, delivery, subscription apps): the chain is Ad → Install → Registration → First Deposit/Purchase → Repeat Usage. An install is worthless if it stops there. This requires mobile measurement partners like AppsFlyer, Singular, or Adjust that can follow one user from ad click through to their first in-app action, something standard web analytics simply can’t do.
- E-commerce businesses: the chain is Ad → Visit → Cart → Checkout → Order Confirmed → Repeat Purchase. This is the most directly measurable of the three, since spend and revenue tie together cleanly. It’s also the model most sensitive to something outside the media plan entirely: website speed and checkout friction. No amount of audience precision fixes a slow checkout page.
Across all three, the same principle holds: a lead-gen or offline-conversion business absolutely cannot run performance marketing without a CRM feeding data back into the ad platforms. Without that feedback loop, “success” gets measured at the wrong point in the funnel, and every subsequent decision compounds that mistake.
5. Analytics: Measuring what actually matters
Every business model needs a different scorecard, and using the wrong one is how “successful” campaigns quietly bleed money. A CPM and CTR tell you almost nothing on their own; they only become useful alongside the metrics further down the funnel that are specific to how that business actually makes money:
- Offline/lead-gen: cost per MQL, cost per SQL, MQL-to-SQL rate, cost per lead, and ultimately ROAS calculated against actual closed sales, not form fills.
- E-commerce: add-to-cart rate, checkout completion rate, and ROAS calculated against real revenue.
- Apps: cost per install, install-to-registration rate, registration-to-transaction rate, and customer lifetime value.
We build every reporting dashboard around whichever of these actually maps to revenue for that specific business, and we deliberately exclude vanity metrics that don’t inform a decision. If a number doesn’t change what we’d do next, it doesn’t belong in the report. The two failure modes we see constantly in businesses managing this in-house are wrong conversion tracking (campaigns get judged on numbers that were never accurate to begin with) and mismatched attribution, where the ad platform, the analytics tool, and the actual sales ledger all tell a slightly different story, and nobody has reconciled which one to trust.
6. AI: Where it genuinely changes the work, and where it doesn’t
AI now touches nearly every stage above it, but the honest answer is that it’s an accelerant on a sound process, not a replacement for one. Inside the auction, machine-learning bidding (Performance Max, Advantage+, and similar tools) can now test and optimize far faster than manual bidding ever could, provided the conversion data feeding it is accurate. That last condition matters more than the tool itself; feeding a bidding algorithm on bad conversion data just means it optimizes toward the wrong outcome faster.
On the creative side, AI has meaningfully lowered the cost and turnaround time for testing multiple ad variations, which matters most at the top of the funnel where volume of creative experimentation tends to correlate with performance. On analytics, predictive modeling and cohort analysis (essential for app businesses, where a user’s true value only becomes clear weeks after install) are dramatically faster to run than they were even two years ago.
What hasn’t changed is the order of operations. AI can make each of the six steps faster and cheaper to execute. It cannot substitute for actually knowing your audience, or for connecting a CRM to your ad account, or for deciding what “success” means for your specific business model. Every client conversation where AI comes up, our answer is the same: simplify the tool, don’t let the tool set the strategy.
Why the framework matters more than any single platform
The pattern across every example above, an industrial pump manufacturer, a real estate developer, a coffee brand, a gaming app, is the same: the platforms and tactics change completely, but the six checkpoints don’t. Audience work looks different for a B2B distributor network than it does for a D2C coffee brand, but skipping it produces the same result either way: expensive clicks that don’t convert.
That’s the actual value of running campaigns against a fixed framework instead of platform-by-platform instinct. It forces every account to get checked at every link, in the same order, every time, rather than discovering after the budget is spent which link quietly broke.
Curious where your business’s performance marketing would benefit from this process? Get in touch and we’ll walk you through it.