Platform · Lead routing
SAGARIS routes accounts, and your assignments stand
Territory rules pick the owner the moment a contact is created, imported or merged. New leads at that company inherit that owner automatically. The write only ever fills an empty owner, so the rep you put on an account keeps it.
Included at $499 per seat per month
How an account finds its rep
priority order, first match wins
name in your named list
employee_count gte 1000
location contains boston
employee_count between 50 and 999
When no rule matches
Territories are checked lowest priority number first, and the first rule that matches ends it. Nothing below the match is looked at. Put a specific territory above a broad one and it always wins, which makes the order on screen the order in real life.
How it runs
Accounts route on their own. Leads route the way you approve.
Two things here assign work, and you get to use both. Territory rules fire on their own, inside the same request that created the record, so an account never waits for anyone. Lead level distribution hands your manager a plan and waits for the nod. One runs itself, one runs on your say-so.
automatic
Territory rules, at the account grain
Runs inside the write that created the record, on eight paths: a contact created, updated or merged, a CSV import, the public API, and three campaign audience imports. The account gets an owner and a territory, and every lead at that company inherits the owner when it is created.
preview, then apply
Round robin and least loaded, at the lead grain
You get the final say on this one. A manager asks for a plan across up to fifty unassigned leads, sees which rep each one would go to and why, and applies the whole thing in a click. Nothing moves until you approve it, so the desk you built stays the desk you built.
Then it ranks what landed
The score is stored on the contact and the database only ever accepts 0 to 100, so the number on screen always means the same thing. Speed to lead is measured against a five minute target and reported in its own view, with tiers at five minutes, thirty minutes and twenty four hours. You can see where the clock is being lost and coach against it this week.
One hundred points, five ways to earn them
hot at 75
- Email engagementmax 25
- Company fitmax 25
- Behavioural signalsmax 20
- Decision-maker levelmax 15
- Recencymax 15
A perfect engagement record on its own is worth 25, which is cold. Hot needs three factors to agree. Each factor is capped, so nothing reaches the top of a rep's list on one loud signal and your best hour goes to your best lead.
A score that ages on purpose
half life, 30 days
Thirty quiet days halve the score, sixty quarter it. So the top of your queue holds leads that are warm right now, not leads that were warm last quarter. The ranking keeps meaning something without anyone tidying it.
Why this is different
Your reps keep the accounts you gave them.
Assignment here is a guarded write that can only ever fill an empty owner. So you can run the whole sweep as often as you like and every decision a person already made survives it untouched. No rep loses an account to a rule overnight, and nobody spends Monday morning putting ownership back.
When the rule is the authority
An owner somebody set by hand gets replaced, and it falls to a person to notice and put it back.
SAGARIS
The person is the authority. An account that already has an owner is not a row the write can match, so the rule passes over it and your call stands.
update accounts set owner_id where owner_id is null
That clause is the whole promise. It travels inside the statement rather than running first, so it holds even when two writes land at once and re-running the sweep is a non-event.
A guard, not a good intention
The condition travels inside the update rather than running before it. There is no window between checking that an account is unowned and writing to it, which is why a manual assignment made at the same instant cannot lose the race.
Safe to run again, and again
Sweep the unassigned queue as often as you like. Anything already owned is skipped, and the round robin cursor advances by how far the batch moved instead of being overwritten, so concurrent runs add up rather than collide.
Ownership lives on the company
Once an account belongs to a rep, every lead at that company inherits that owner as it is created. Two reps do not end up working the same building because two enquiries arrived through different doors on different days.
The questions buyers actually ask.
- MOST ASKED
Territory routing is. Rules run on eight write paths. They cover a contact created in the app or through the public API, an update, a merge and a CSV import. Three campaign audience imports are covered too. The account is assigned inside the same request. Round robin and least loaded work differently on purpose. They produce a plan a manager reads and applies, so nobody is handed accounts by a rule they never saw.
Manual assignments hold. The update can only match accounts whose owner is still empty, so an account a manager assigned is passed over rather than corrected. The guard is part of the database statement itself, not a check run just before it. It still holds when a rule sweep and a manual assignment land at the same moment.
Five rule types, combined with six operators, which is what lets a territory be described without writing code. The types are geographic, industry, company size, revenue tier and custom; the operators are equals, contains, in, greater than or equal, less than or equal, and between. Territories are evaluated in the priority order an admin sets, lowest number first, and the first rule that matches ends the search, so a specific territory placed above a broad one takes the account.
It waits somewhere visible rather than being guessed at. By default an unmatched lead goes to an unassigned queue instead of being handed to whoever happens to be next in the rotation, because a silent guess about ownership is worse than an obvious gap somebody can see and fix. A workspace can opt into a round robin across the territory owners instead, and the cursor for that rotation is persisted and advanced atomically, so the next batch continues where the last one stopped rather than restarting at the first rep.
Five factors that add up to one hundred: email engagement up to 25, company fit up to 25, behavioural signals up to 20, decision maker level up to 15 and recency up to 15. Title is worth 15 for c-level, 12 for a VP, 9 for a director, 5 for a manager and 2 for an individual contributor. Recency pays 15 for activity inside a day, 12 within three days, 8 within a week, 4 within a fortnight and 2 within a month. Seventy five and above is hot, 50 is warm, 25 is cold, below that is ice, and the number is stored on the contact behind a database constraint that will not accept anything outside 0 to 100.
Scores decay, which is what keeps the queue worth working. A lead that scored 80 and then went quiet for a month reads 40, and 20 a month after that. The half life is thirty days. Without that, the top of the list slowly fills with people who were interested last quarter and your reps stop trusting it.
Territories and their order are yours. You write the rules and set the order they are checked in. The first match wins, so a specific territory placed above a broad one takes the account. You choose whether an account no rule matched waits in a visible queue or rotates round robin across the territory owners. Scoring runs on five fixed factors, so a score of 80 means the same thing in every workspace and every team.
Bring your territories. We will run them on your accounts.
Bring the rules you use today and a real slice of your accounts. We will show you which territory each one lands in, which rule matched it, and which accounts nothing matched at all. You walk away with your own coverage map and a clear list of the gaps worth closing first.
In this topic