How agencies prove the leads they deliver get handled
Agencies prove the leads they deliver get handled by reviewing the conversations those leads landed in, not by counting how many they sent. That review is Sales QA: a graded pass over the client's Instagram, Facebook, and WhatsApp threads, with the original conversations attached to every claim.
The scope line nobody writes down is this. You are paid to produce conversations. The client is responsible for having them. Your reporting stops at the first of those, the disagreement starts at the second, and an unwritten line defaults to whoever pays.
Why can nobody win the argument about lead quality?
Because both sides are describing the same conversations and neither of them has opened one.
Their evidence is what the sales team said in a Monday meeting: these are tire kickers, they never buy. Yours is conversations started and what each one cost. Both are proxies for what happened inside the threads. So the meeting goes to whoever sounds more certain, and over enough meetings to the client, because the retainer is theirs.
Your numbers also lose that argument when they are right, for a structural reason: they are your own report card. And bad is a conclusion, not an observation. It is what a business says when revenue did not move and the only visible input is your invoice. Paid reporting stops at the first message, which is where ad instrumentation ends, and exactly where your accountability is assumed to continue.
What does your monthly report prove, and what is the client asking?
Your report proves your half ran. Spend, reach, cost per result, leads delivered, a chart that points up. Every line is true, and every line stops where the client's business takes over.
The client is asking one thing, did anything come of this. The test for a line is whether they can act on it on Monday.
| The line you send today | The line that replaces it | What the client can do with it |
|---|---|---|
| Leads delivered this month | Leads delivered, and how many asked about price, availability, or booking | Settle whether the argument is about targeting or about handling |
| Reach, impressions, engagement rate | Nothing. Move them to an appendix | Stop spending meeting minutes on numbers nobody acts on |
| Leads have been passed to your team | Where those conversations stopped, grouped into named patterns with counts | Fix one script instead of relitigating lead quality |
| Recommendation: increase budget | Recommendation: one sentence to add after every price answer | Make a change that costs nothing and shows in the next report |
Keep the two halves in separate sections. Nothing signals an excuse faster than a slide where your spend and their reply times share an axis.
Which evidence settles each claim about lead quality?
Every claim in that meeting is a count with threads behind it.
| The claim | Who makes it | The evidence that settles it |
|---|---|---|
| The leads are bad, none of these people are serious | Client | How many opening messages asked about price, availability, or a booking |
| Your team never answered them | Agency | Threads where a person wrote back, against threads that got only an automated acknowledgment |
| We answered every single one | Client | Time to that first human reply, split into working hours and after hours |
| They answered and never asked for the sale | Agency | Whether the last business message names a time, an action, or a question |
| These people message at midnight, that is not on us | Client | The hour the stalled leads arrived, next to reply times in those hours |
No row holds an opinion, a rating, or a name. Note who each row convicts: rows two and four land on the client's sales floor, rows one and five land on you.
What if the leads really were bad?
Then the same review says so, and you say it first, before the client's sales lead gets there.
An agency that only ever produces evidence blaming the client is running a defense, not a service, and clients work that out. The tell is arithmetic: if the handled share is healthy and the share of conversations carrying a buying question is low, the process is fine and the demand is not. That one is yours, and its fix is as concrete as the ones you hand them: move the qualifying question into the ad instead of the DM.
What you cannot do is file the unanswered ones as somebody else's problem for another month. In the Sprout Social Index, 2025, 73% of social users said that when a brand does not respond to them on social, they will buy from a competitor instead. You delivered and billed those threads. In your report they are delivered. To those customers they are a business that ignored them.
How do you raise this without blaming the client for your numbers?
Lead with your own half, bring counts instead of names, and finish with a decision they can make in the room.
- Present the review at a fixed point in the month, never in the renewal meeting. Evidence produced while you ask for budget reads as leverage.
- Bring counts, not names. Their staff are not yours to grade, so the discipline that keeps a finding from becoming a verdict on somebody applies double to a vendor.
- Hand over the threads. A review the client cannot browse is only a better-formatted claim.
The example Sales QA report on the homepage shows the finished shape of that second half. One sample week: 420 conversations checked, 96 potential buyers, 74% replied or handled, 9 still waiting, and first replies averaging 41m in working hours against 1h 23m outside them. Price came up in 12 conversations and 7 stalled. That is example data from one sample account, not research and not a benchmark.
Four sentences that work in a real meeting, worded on those example figures. None asks anybody to admit they are bad at their job.
- Our first messages promise a same-day callback and the replies run longer. It is our copy before it is your staffing, and we are rewriting it this week.
- Nine are still waiting. I would rather fix that than send you more.
- Twelve of these asked what it costs and seven stopped at the answer. That is one missing line in a script, and it is cheaper than budget.
- The evening is the biggest single gap. I do not need it fixed today, I need to know who owns it.
The sentence that costs you the account is your team dropped the ball. It asks the person paying you to agree out loud that they run a bad shop.
Who is allowed to read a client's conversations?
Only the people the client granted access to. Access is granted, never assumed.
In Reploom, workspace access and organization billing access are separate. An Organization Admin can use available plan capacity to create workspaces, but does not automatically gain access to the conversations inside them. A Workspace Owner or Admin manages integrations, teammates, workspace settings, and the Sales QA controls. Billing belongs to the organization, not to an individual user.
The shape that falls out is one client per workspace, and one conversation you owe them at kickoff: their customers' messages will be read for review, by named people on your side, and they agree to it in writing. That is their call, not a line item in your scope. A client who finds out in month three is right to be annoyed.
One onboarding fact. On WhatsApp, monitoring starts at connection, so how far back a review reaches depends on available history. With a client you already work with, the first weeks are your baseline, not the back catalog.
What Reploom does for an agency, and what it will not do
Reploom is AI Sales QA for social conversations. It reviews Instagram, Facebook, and WhatsApp messages, comments, and mentions after they happen, scores them for buying intent, and groups repeated misses into named patterns in a weekly report with the source conversations as evidence. It does not draft, send, or approve replies, it is not an inbox or a chatbot, and it does not replace the people answering DMs.
It is not an agency platform either. There is no white-label mode, no client portal, and no branded export: the report you send is yours to write, which is the part clients pay you for. It does not connect to any ad platform, does not calculate attribution, and cannot tell you which campaign a thread came from, so setting spend beside conversation quality stays arithmetic you do by hand.
The meeting changes once the threads are on the table. It gets shorter. Nobody defends a feeling, somebody picks up the evening, and the rest of the hour goes to the two things you each agreed to change. Clients rarely renegotiate in the middle of a list they helped build. Book a Demo to see a client's month reviewed that way.