Skip to content
HomeStar Docs

Why Archive Isn't Terminal

In most CRMs, archive means dead. The lead is filtered out of every active view, the drips are stopped, the agent moves on. The system implicitly says: this person is no longer in your pipeline; they may as well not exist.

ARCHIVED in HomeStar is not that. It’s a holding state — a place where leads who’ve stopped engaging wait, often months, sometimes longer, until the system or an agent notices they’ve come back. The re-entry edge from ARCHIVED → CONTACTING is a first-class transition, not an exception.

This page is the argument for that design choice.

What “terminal” means in a state machine

Section titled “What “terminal” means in a state machine”

A terminal state has no outgoing transitions. Once a thing reaches a terminal state, the state machine is done with it. CLOSED is the closest thing the pipeline has to a true terminal — the deal funded, the keys exchanged, the relationship transitions out of the active-prospect mode.

ARCHIVED looks terminal at first glance. The lead is set aside. Their card is in the rightmost column. They’re not in your daily attention.

But they have a way out: ARCHIVED → CONTACTING, triggered by lead_re_engaged. That transition fires when:

  • The system sees the lead come back to the site and engage (property views, favorites, returned visits push the temperature score up).
  • Or an agent manually reactivates them.

That single edge makes ARCHIVED recoverable. The choice to include it changes the meaning of the column.

The “leads come back” problem

Section titled “The “leads come back” problem”

The empirical case for the re-entry edge: real estate leads come back more often than agents migrating from other CRMs expect.

Picture a lead who inquired about a property six months ago. They were pre-approved but the listing went under contract before they could see it. You followed up; they went quiet. After 21 days of no response, the system archived them.

Six months later, they’re looking again. They remember your brokerage and come back to the site. They view three listings, save a favorite, run a search.

Two questions:

  1. Does your CRM tell you they’re back?
  2. If you knew, would you reach out?

In a system that treats archive as terminal, the answer to (1) is no. The lead is filtered out of every active view. The temperature scorer might still be running, but there’s no surface where you’d see the score climb. The behavior you most want to know about — they’re shopping again — is invisible.

In HomeStar, that lead is still in the ARCHIVED column on the board, the temperature chip on their card is now Warm or Hot, and you see them when you scan the board. The system makes the return visible because the column isn’t a graveyard — it’s a holding area with the lights on.

What ARCHIVED is for, then

Section titled “What ARCHIVED is for, then”

If it’s not a dead-lead bucket, what is it?

A few characterizations that all point the same direction:

  • A scheduled-quiet column. The lead was working before and has gone unresponsive long enough that continuing daily attention is wasted effort. Setting them aside lets you focus on leads who are actively in motion.
  • A long-tail watchlist. The system continues to observe behavioral signals on archived leads. Most of those signals are nothing — the lead doesn’t return. The few that matter — return visits with concrete intent — get surfaced by the temperature chip on the card.
  • A history-preserving alternative to delete. Deleting the lead loses the conversation, the prior context, the prior trust. Archiving keeps all of it intact so that if the lead comes back, the next outreach starts from “I remember you, you were looking on the West End,” not from “hello, this is a cold call.”

The principled framing: ARCHIVED is the column for leads not currently in motion. The state machine assumes some of them will start moving again.

The Lawrence Pickett scenario

Section titled “The Lawrence Pickett scenario”

In the demo, Lawrence Pickett is currently in the NEW column with a breached SLA and a DNC flag. He’s the case study for the breach-and-recover pattern in your first day.

Now imagine Lawrence two months from now. You and your team made contact, worked the conversation for a few weeks, but after the DNC was sorted out and the initial enthusiasm faded he stopped responding. The 21-day timeout fires. Lawrence goes to ARCHIVED.

Four months later, Lawrence’s situation has changed. He took a new job, his wife is expecting, they need a bigger place. He thinks about the brokerage he talked to last year and comes back to the site. He views six listings in one evening, saves three favorites, fills out an inquiry on a different property.

In a system that treats archive as terminal, that inquiry would land as a brand-new lead — no history, no context, possibly routed to a different agent. The fact that you’d spoken to Lawrence before would be invisible. The CRM would be telling you “new lead!” when the right framing is “Lawrence is back.”

In HomeStar, the system detects Lawrence’s recent activity, the temperature on his archived card climbs, and his existing record gets pulled back into CONTACTING via lead_re_engaged. You see the original timeline — your prior conversation, the original archive — and the new signals.

The continuity of identity is the win. Re-entry is just the mechanism that delivers it.

What this design costs you

Section titled “What this design costs you”

Worth being honest about the tradeoffs:

  • The ARCHIVED column is never empty. Old leads accumulate. The board doesn’t auto-prune. If you find this visually noisy, sort or filter the column when reviewing.
  • Temperature scoring runs on archived leads too. That’s a small ongoing compute cost. It’s also why the re-entry signal works.
  • A lead can re-enter your queue when you weren’t expecting them. This is the point, but it does mean your day can include “Lawrence is back, call him” as a surprise item.
  • Genuine dead leads stay in the dataset. If a lead truly never comes back, they sit in ARCHIVED forever. The cost of keeping them is approximately zero; the cost of deleting them and then finding out they came back is high.

When archive is actually the wrong column

Section titled “When archive is actually the wrong column”

Archive is for real people whose engagement has stalled. It’s not the right column when:

  • The contact data is bad — that’s TRASHED, and TRASHED has its own re-entry path via contact_info_updated.
  • The lead has actively asked not to be contacted — set the appropriate communication flag (dnc or a channel-specific one); archive on top of that is fine but the flag is what does the protection.
  • The lead has explicitly told you they bought elsewhere or stopped looking — that’s still archive, but a manual one with a clear note, not a timeout.

The takeaway

Section titled “The takeaway”

Archive in HomeStar is a holding state. The re-entry edge from ARCHIVED → CONTACTING is the design choice that makes the column meaningful rather than just a delete-with-extra-steps. The thing it protects against is silent loss — a lead who came back, was ready to buy, and never made it onto your radar because the CRM had given up on them.

The cost is a fuller archive column and the occasional surprise re-entry. The benefit is that you actually know when a lead is back.

Section titled “Related”