Rebuilding the Lead Lifecycle When Sales and Marketing Disagree

Rebuilding a lead lifecycle model is really just couples therapy for marketing and sales, except the couple has a CRM instead of a shared bank account and everyone's mad about attribution instead of dishes. The fix is a shared model with stages both teams agree to, thresholds that get reviewed instead of worshipped, and a handoff that's instrumented so nobody has to trust anybody's gut. Here's the setup you've lived through in some form. Marketing says "we sent you 400 MQLs this quarter." Sales says "we got maybe 40 leads worth calling, and half of them worked at companies that don't exist anymore." Both sides have a spreadsheet. Both spreadsheets are technically correct. Neither spreadsheet agrees with the other, and that gap is where quota gets missed and blame gets assigned. The root cause almost never has anything to do with lead quality. It's a definitions problem wearing a data problem's clothes. ### Why "MQL" Became a Swear Word Marketing Qualified Lead used to mean something specific: a lead that hit certain behavioral or firmographic signals and got handed to sales. Somewhere along the way, every team built its own private dictionary. Marketing's MQL might mean "downloaded a whitepaper and works at a company with more than 50 employees." Sales' definition of a real lead might mean "someone who's actually said, out loud, that they have budget." Those are not the same lead. One is a temperature check. The other is a pulse. When those definitions live in two different heads instead of one shared doc, you get the classic fight: marketing hits its number, sales misses its number, and the CRM sits there full of "MQL" tags that mean absolutely nothing because everyone stopped trusting the tag six months ago. At that point the label is decoration, a participation trophy stapled to a lead record. I once sat in on a QBR where a VP of Marketing proudly announced a 40% increase in MQLs. The VP of Sales just stared at his coffee and said, "Congratulations, you've found a more efficient way to generate leads we're not going to call." Nobody laughed, because everyone in the room knew he wasn't joking. That's the moment the company finally agreed to rebuild the model from scratch, because a metric that makes one team look good and the other team look lazy is a grenade with the pin already pulled. ### Step One: Stop Arguing About MQL and Start Mapping Stages Before anyone touches a scoring model, get both teams in a room (or a Zoom, no judgment) and map the actual lifecycle, stage by stage, with plain language everyone can defend. Something like: - Subscriber: gave you an email, nothing more. Barely a relationship. It's a first date where they haven't even told you their last name. - MQL: showed real engagement signals, matches your ideal customer profile on paper. - SAL (Sales Accepted Lead): sales looked at it and agreed it's worth their time. This stage is the whole point of this article, so don't skip it. - SQL (Sales Qualified Lead): sales has talked to them, confirmed budget, authority, need, timeline, or whatever your version of BANT looks like. - Opportunity: it's in the pipeline, there's a dollar figure attached, and someone's forecasting against it. Notice SAL sits between MQL and SQL. That stage is the missing piece in most broken models, because it's the one moment where sales gets to say "yes, I accept this" or "no, send it back" before it counts against anyone's numbers. Without it, marketing scores a lead as qualified and sales either ignores it or works it and calls it garbage. Either way, nobody agrees on what happened. ### Setting Thresholds Without Guessing Scoring thresholds fail for one boring reason most of the time: someone picked a number that felt right and never touched it again. 75 points feels authoritative, but it's arbitrary unless you've done the work to back it up. The better approach is to look backwards before you look forwards. Pull your last two or three quarters of closed-won deals and reverse-engineer what those accounts and contacts actually looked like at the MQL stage. What job titles converted? What company sizes? What behaviors (demo requests, pricing page visits, three or more email opens in a week) actually showed up before deals closed, versus what showed up on leads that went nowhere? You'll usually find your scoring model has been rewarding the wrong signals. Newsletter opens get treated the same as pricing page visits in a lot of legacy models, which is like giving equal credit to someone who waved at you across a bar and someone who asked for your number. Fix the weighting so behavioral signals that correlate with actual buying intent (pricing pages, demo requests, competitor comparison pages) outweigh passive engagement (opens, single-page visits). Then set the threshold at the point where conversion rate to SQL starts climbing meaningfully, rather than the point where lead volume looks impressive on a slide. Volume without conversion is a vanity number, confetti. ### Instrumenting the Handoff So Nobody Has to Take Anyone's Word for It This is the part that actually lives in the CRM, and it's where most rebuilds either succeed or quietly rot within two quarters. **Build the SAL stage as a real, required step,** with an actual decision point where a sales rep or SDR has to click "accept" or "reject" within a set window, say 24 or 48 hours. If it sits untouched past that window, it should auto-escalate to a manager. Silence shouldn't count as acceptance, and it definitely shouldn't count as rejection either. **Log the rejection reason every time.** "Bad fit," "wrong title," "already a customer," "duplicate," whatever it is; capture it as a picklist, not a free-text field nobody reads. After a quarter of this data, you'll know exactly which lead sources are quietly feeding sales garbage and which ones are gold. That's the whole point of instrumenting the handoff. It turns a shouting match into a report. **Set up lifecycle stage automation with guardrails, not just triggers.** A lead becomes an MQL when it crosses the score threshold, fine, that part's easy. But don't let it silently become an SQL just because a rep changed a status field on a Friday afternoon trying to clean up their pipeline view. Stage changes tied to revenue should require the actual qualifying action (a scheduled call, a confirmed budget conversation) not a dropdown click. **Give both teams the same dashboard.** One shared view of MQL volume, SAL acceptance rate, SQL conversion rate, and time-in-stage at each step, rather than separate marketing and sales dashboards that pull from different reports and mysteriously never match. When both teams are staring at the same numbers, the fight changes shape. It stops being "your data is wrong" and starts being "our acceptance rate dropped in March, what happened." That's a solvable problem. The other one isn't. Quick gut check, and you already know the answer: what happens to a lead lifecycle model that nobody reviews after launch? It rots, the same way it did before you rebuilt it, because a dashboard nobody opens is just expensive wallpaper. ### The Review Cadence Everyone Skips Here's the part nobody wants to hear: this model needs a checkup every quarter, not a one-time rebuild you high-five over and never revisit. Buyer behavior shifts, your product changes, your ICP drifts as you move up or down market. A scoring model built for a $5,000 deal size doesn't necessarily hold up once you're selling $50,000 contracts to a different buyer persona entirely. Put a recurring meeting on the calendar. Pull the SAL rejection reasons, the conversion rates by stage, the time-in-stage numbers. Ask the boring but necessary question: is this threshold still doing its job, or are we just used to it? Thresholds that never get revisited aren't standards anymore. They're superstitions. ### The Part Where Everyone Has to Actually Agree None of this works if it's marketing's model that sales tolerates, or sales' model that marketing works around. It has to be built together, signed off by both sides, and revisited together. That sounds like a soft, feelings-forward ask for two teams that mostly communicate through Slack threads with increasingly passive-aggressive punctuation, but it's an operating agreement, and it needs the same rigor you'd put into a contract, because functionally, that's what it is. Get the stage definitions in writing. Get the thresholds tied to actual conversion data, not vibes. Instrument the SAL handoff so acceptance and rejection are visible, timestamped, and reasoned. Do that, and the argument about what counts as an MQL mostly disappears, not because everyone suddenly agrees, but because there's finally a scoreboard both teams trust enough to stop arguing about the score.