Gully Sales builds a scoring model from what your own won deals have in common, so every lead arrives ranked, with the reason for its rank visible to the person who has to act on it.
Gully Sales Private Limited works with small and medium businesses across India, inside the CRM your team already opens every day.
In one paragraph
Lead scoring gives every lead a number that says how closely it matches the customers you already win, and how much interest it has shown recently. Gully Sales builds that model from your own closed deals, writes the thresholds that move a lead to sales, configures it inside your CRM, and reviews it against results.
The problem
Most growing businesses do not have a lead problem. They have a lead order problem. Enquiries arrive from the website, from ads, from exhibitions, from referrals and from WhatsApp, and they all land in the same list looking equally important. The person who opens that list has no way of knowing which enquiry resembles the customers you already serve well, so they start at the top and work down. The strong lead sitting at number forty waits its turn.
Why it persists. Scoring feels like a reporting job, so it never reaches the top of anyone's list. The people who could define it are busy selling. When it is attempted, it is usually copied from a template built for another market, gives almost every lead a similar number, and quietly stops being used within a month. And nobody enjoys saying out loud that one lead is worth less than another, so the rating field stays polite and empty.
If it stays unresolved. The gap widens as volume grows. More enquiries mean more of your strongest ones buried further down the list, and more hours spent proving to each other that the fault lies with the other team. Businesses in this position often respond by buying more leads, which makes the queue longer without making it any clearer.
What changes
In the first weeks
In how the work runs
In sales and marketing
In what management can see
Over the longer term
Gully Sales controls the model, the rules, the build inside your CRM and the review that keeps it honest. Whether revenue follows depends on your offer, your pricing, your capacity to follow up and market conditions. A score tells your team where to spend attention; it does not close the deal.
Who it is for
What Gully Sales does
Ideal customer fit scoring
We take the customers you have won and kept, and separate the traits they share: industry, size, location, buying role, order pattern, and the problem they arrived with. Those traits become the fit half of the score, so a lead is measured against your actual customer base rather than against a borrowed template.
Behaviour and intent signals
We map the actions a lead takes before buying: pages viewed, forms completed, emails opened, quotes requested, calls answered, WhatsApp replies. Each is scored by how often it appeared before a real deal. Weak signals stay weak, and the actions that genuinely precede a purchase carry the weight.
Score bands and thresholds
A number on its own changes nothing. We agree the bands: the score at which a lead becomes sales-ready, the range that stays in nurture, and the level at which a lead is set aside. Then we write what each band obliges somebody to do, and by when.
Recency, decay and re-entry rules
Scores go stale. We set how quickly behaviour points fade when a lead goes quiet, when a lead re-enters scoring after a period of silence, and how a returning contact is treated. Fit points hold steady; interest points decay on a stated schedule.
Exceptions and manual overrides
Some leads matter for reasons no model can see: a referral from your largest customer, a name from a target account list, a tender you have waited two years for. We define the exceptions that bypass the score, who is allowed to apply them, and how each override is recorded.
Scoring built inside your CRM
We configure the model where your team already works: the fields, the calculation, the score on the lead record, the reason text explaining the number, and the list views and alerts that put high scores in front of the right person. There is no new tool to open.
Calibration against closed deals
We test the model against deals already closed. Did the leads that became customers actually score highly, and did high scores convert more often than low ones? Where they did not, we change the weights rather than defend them.
Score health monitoring
We set up the reporting that shows whether the model still works: the spread of scores across new leads, conversion by band, the share of leads sitting above the sales threshold, and how often overrides are being used.
How it runs
Baseline and data check
We look at how leads reach you today, which fields are filled reliably, how many enquiries arrive in a typical month, and how many closed deals we have to learn from. If the data cannot carry a score yet, we say so before any model is designed.
Won and lost analysis
We separate the customers you won and kept from the deals that stalled or were lost, then look for the traits and actions that distinguish them. This is where the model gets its opinions, and they come from your history rather than from a generic framework.
Model design workshop
We put the draft in front of sales and marketing together, argue the weights, and settle what a fit point and a behaviour point are each worth. Disagreements here are useful. They are the same disagreements that otherwise surface later as complaints about lead quality.
Thresholds, ownership and response
We fix the score at which a lead becomes sales-ready, who receives it, how quickly it should be answered, and what happens when nobody answers. Exceptions and overrides are defined here too, along with the people allowed to use them.
Build and configuration
We configure the model inside your CRM: the fields, the calculation, the reason text, the views your team works from, and the alerts for high scores. We keep it readable, so anyone can see how a number was arrived at.
Pilot and calibration
The model runs alongside existing habits for an agreed stretch. We compare the scores it produces against what the team judges and what the deals do, then adjust weights and thresholds while everyone is still paying attention.
Handover and training
We train the people who work leads and the person who will maintain the model. The training covers reading a score, using an override honestly, and knowing what to escalate when the list stops looking right.
Review cycle
The model is reviewed on an agreed rhythm against fresh outcomes. Markets shift, offers change and new sources appear, and each of those moves the weights. Review is what keeps a model useful after its first quarter.
Ways to work with us
Scoring model design
A focused engagement to analyse your won and lost deals, design the model, agree bands and thresholds, and hand you a specification your own CRM team can build.
Scoring model design and build
The model designed and then configured inside your CRM, including fields, calculations, views, alerts and the sample records that show your team how a score reads.
Scoring within a revenue operations engagement
Lead scoring built alongside capture, routing and lifecycle work, for businesses where the underlying records and handover rules need attention at the same time.
Recalibration and review retainer
A standing review of a model you already run: calibration against recent outcomes, threshold adjustment and reporting on score health, without rebuilding what works.
Why Gully Sales
We build the model from your deals, not from a template.
The attributes and signals come from customers you have already won. A model borrowed from another market scores confidently and wrongly, and your team stops trusting it by the second month.
We work inside the CRM you already have.
Gully Sales configures scoring in the system your team opens every day. There is no separate tool to log into, and no reporting layer that only one person in the office knows how to run.
We treat sales and marketing as one problem.
Scoring fails when one side designs it. We take the weights, thresholds and handover rules through both teams together, so the number carries an agreement rather than a preference.
We keep the model readable.
Every score carries the reason behind it in words. A seller can see why a lead ranks where it does, disagree with it out loud, and give us something specific to correct at the next review.
We plan for the model to age.
Markets move and offers change. Calibration and monitoring are built in from the start, and the person who owns the review is named, so the model does not quietly become decoration.
Where it applies
Manufacturing
Clinics and healthcare
Professional services
Interior design and real estate
Education and training
Proof
Premier Marketing
Questions buyers ask
Two things do it together. The score identifies the lead, and the band rules say what must happen next. A lead crossing the sales-ready threshold triggers an alert to the owner named for that band, appears at the top of a prioritised list view, and carries the reason for its score on the record. The person receiving it does not have to hunt for context or decide for themselves whether it matters.
It depends on how many sources you run, how complete your CRM records are, and how many closed deals we have to learn from. A business with one clean source and two years of history moves quickly. A business with five sources and half-filled records spends time on data first. We give you a schedule after the readiness check rather than before it, because a date set without knowing the data is a date that slips.
Access to your CRM or a clean export of your leads, a list of every source feeding it, your won and lost deals for as far back as they exist, and time with the people who actually close business. The last one matters most. The traits separating a good customer from a poor one usually live in a seller's head long before they live in a field on a record.
By whether high scores convert better than low ones. We test that against your own outcomes: conversion by band, response speed above the threshold, and how often overrides are used, which shows where the model disagrees with your team. If band and outcome do not move together, the weights are wrong and we change them. The model is judged on results, not on how sensible it looked at design.
We do not generate leads, run campaigns or make calls on your behalf. Scoring ranks the leads you already receive. We also do not rebuild a CRM that is not yet in daily use, migrate historical records or write your sales scripts as part of this work, though we will tell you plainly when one of those needs to happen first, and we can take it on separately.
They work together. Scoring is automatic and happens before anyone speaks to the lead: it ranks every enquiry using data already on the record. Qualification is a conversation, where a person confirms need, budget, authority and timing. A score decides who gets called first. Qualification decides whether the lead becomes an opportunity. Scoring without qualification ranks strangers; qualification without scoring simply starts at the top of the list.
You need one place where every lead is recorded and kept current. That is usually a CRM, and it has to be in genuine daily use rather than merely installed. If leads live across inboxes, spreadsheets and personal phones, a score built on them will be confidently wrong. When that is the situation we say so at the readiness check, and deal with capture and records first.
No. A low score sends a lead to nurture, not to the bin. Those leads keep receiving communication and keep being scored, so a contact who becomes active later rises on their own without anyone having to remember them. The bands state what each level obliges someone to do, including the lowest one. Nothing is dropped silently, and every lead has a stated destination.
The model we build here is explicit and rule-based: named attributes, named signals, stated weights, and a reason written in words on every record. Your team can read it, argue with it and correct it. Predictive scoring that learns patterns from data is separate work with its own data requirements, and we treat it that way rather than presenting a black box and calling it a scoring model.
Fewer than most teams expect. A model with eight to twelve well-chosen attributes and signals usually ranks better than one with forty, because each additional weak signal pulls scores towards the middle until everything looks average. We include a signal only when your own closed deals show it mattered, and at each review we remove the ones that have stopped earning their place.
Someone in your business, named at handover, usually the person who administers the CRM or runs marketing operations. They receive the model document, the configuration and a maintenance guide, and they run the review if we are not retained for it. A model with no named owner drifts, because nobody has the standing to change a weight when the market moves.
That is useful, and we design for it. Sellers can override the score with the reason recorded, and we read those reasons at every review. A pattern of overrides in one direction is the clearest evidence that a weight is wrong. The alternative, a model nobody is allowed to argue with, gets ignored quietly instead, and you find that out a quarter later.
Talk to us
The first conversation is a discussion, not a pitch. We look at your lead sources, your CRM and your closed deals, and tell you honestly whether a scoring model would help you now or whether something else should come first.