Why follow-up gets forgotten
Follow a W-9 request by checking its status yourself or receiving automatic notifications when its status changes. The form goes out, the vendor or contractor is asked for a name, a tax classification, and a taxpayer identification number, and the team moves on to the next task. Weeks later, someone asks why a payee's file is still empty. Nobody can say whether the form was never opened, was started and dropped, or went back to an inbox that nobody checks.
In a small accounting firm, this gap usually shows up at the worst time, when a staff accountant needs a complete file and finds a hole. The request may have been sent by one person, the reply may land in another person's inbox, and the payee record may belong to a third. Nobody is doing anything wrong on purpose. The process simply has no owner at each stage.
That gap is what this post covers. Once a request is out, a firm has two practical ways to learn what happened. It can check the status itself, or it can have changes reported to it automatically. Both work. They suit different teams, and they fail in different ways. The operations lead's job is to pick a default, write it down, and make sure everyone who handles payees knows which one applies.
It helps to define "done" before you choose a method. For most firms, a request is finished when the form is back, complete, and filed with the payee record, and anything unusual has been reviewed. If the team does not agree on that definition, no status tool will fix the problem. Someone will see that a reply came back, assume the job is over, and leave the file unreviewed.
Before you set up either option, know where the requester rules come from. The IRS publishes Instructions for the Requester of Form W-9. It added guidance for disregarded entities completing lines 1 and 2, and it updated the note in line 3a to emphasize the distinction for LLCs. Your staff does not need to memorize those changes. They do need one shared reference, so the person chasing a missing form and the person reviewing it work from the same instructions.
Option one: check the status yourself
The first option is pull-based. Someone on your team looks up the request and reads its current state. Where a collection service offers an API, this is often the way to do it. TaxBandits, for example, documents a Status endpoint that retrieves "the current status of form submission and TIN Matching result (If opted)" ( TaxBandits W-9 API overview). The phrase "if opted" matters. TIN matching appears in the status only when it was switched on for the request, so a status check tells you nothing about TIN matching if nobody turned it on.
Access is the other detail to plan for. The same overview says that all endpoints require authentication through the OAuth 2.0 flow ( TaxBandits W-9 API overview). In plain terms, the person or system checking status needs working credentials set up in advance. If the one person who knows how to authenticate leaves the firm, the status check quietly stops working, and nobody notices until a file is missing. Assign that access on purpose, and write down who holds it and where the setup steps are kept.
The same documentation lists a List call, described as "List all W-9 certificates associated with a specific payer." That is useful for a roster check. Take your current payee list, compare it with the certificates the service returns for that payer, and the gaps appear in one place instead of in scattered emails.
What a status check tells you has limits. It shows what the service records about the request. Treat a status check as an answer to one question, "has it come back?", rather than as proof that the file is ready.
The strength of pulling is control. The team decides when to look, and nothing lands in anyone's inbox that they did not ask for. The weakness is memory. A status check that depends on one person remembering to run it rises and falls with that person's calendar, and calendars change. For a small number of open requests, a short recurring review usually fits. Put it on a schedule your team sets, name an owner, and record each result in the payee file so the next person can see what was checked and when.
A useful review is short. For each open request, the person reviewing should be able to answer three questions: when was it sent, has anything come back, and who is following up. If nothing has come back and the request is older than the team's usual turnaround, it goes onto the follow-up list. The review does not need to be long. It needs to happen every time, and it needs to leave a record.
When a payee has not responded, the first step is a plain reminder that names the request and explains why the firm needs it. Keep the tone neutral and the message brief. If the reminder gets no reply, do not let the request sit. Pass it to the person who manages the relationship with that payee, so someone with context decides what happens next.
Option two: let notifications come to you
The second option is push-based. The service sends a message when a W-9 certificate's status changes, so nobody has to look up that change. The same TaxBandits overview describes webhooks as "an alternate to the Status endpoint." You can configure webhooks "for the event type 'WhCertificate Status Change' to receive automated notifications" ( TaxBandits W-9 API overview).
The benefit is that a change arrives where your team already works. A reported status change does not have to wait for the next scheduled check. This matters most when the number of open requests grows, or when the person who would do the checking is often busy or away. In those cases, the alert does the remembering.
Setup is the cost. Someone has to configure the webhook for the correct event, point it at a receiving location your team controls, and make sure a named person handles whatever arrives. A notification sent to a place nobody watches is just another unread inbox. Test the path once, from the first trigger to the moment a person sees the message, so you know it really arrives and reaches someone who will act on it.
Where the alert lands matters too. A message posted to a shared queue that several people watch is easier to hand off than one sent to a single personal inbox, because the queue keeps working when someone is out. If more than one person can receive alerts, decide in advance who acts on them. Otherwise two people chase the same payee while a third assumes someone else already has it.
Notifications also have a blind spot. A status-change event fires when the status changes. Without a status change, there is no status-change notification to act on. The message itself will not tell you that a request has been sitting for a long time. That is why push and pull work best together. The notification handles the movements you can see, and a periodic look at the open list catches the requests that have gone quiet.
When an alert does arrive, the team still has to close the loop. Read the change, find the payee record, update it, and decide whether anything else is needed. An alert that is read but not recorded is only half of a process.
Keep a record of every alert that arrives, including the routine ones. That log shows whether the setup is working, and it makes a gap easier to spot later. If alerts stop arriving for a while, you can check whether nothing changed or whether the path broke, instead of guessing.
The two options side by side
The table sets the main options next to each other. Use it to decide which one fits your team, then confirm the details against the service's own documentation before you build anything. The two options differ most in what they depend on. One depends on a person acting, and the other depends on a system being set up and watched.
| Comparison | Check status yourself | Receive notifications |
|---|---|---|
| How it works | Someone looks up the request and reads its current status | The service sends a message when a W-9 certificate's status changes |
| Named example in the TaxBandits overview | The Status endpoint | Webhooks for the event type "WhCertificate Status Change" |
| What you learn | Form submission status, and TIN Matching result if it was opted in | That a status change happened |
| Setup needed | Working authentication and a named checker | A configured webhook, a receiving location, and a named handler |
| Main risk | It happens only when someone remembers to check | No change means no message, so silence can hide a stalled request |
| Best fit (our view) | A small number of open requests and a regular review slot | A larger number of requests, or a team that cannot rely on memory |
Choosing one and writing it down
Most teams should not treat this as either-or. Choose a default based on what your service actually offers and how your team works. If notifications are available, set them up and test them, then keep a short review of open requests on the calendar. If notifications are not available, make the status check a named recurring task with an owner, and do not let it live in one person's head.
Also check that the option you pick matches what your payees will experience. A payee may receive the request under the firm's name even when a service sends it, and may expect the firm to answer questions about it. Make sure someone at the firm can answer those questions, even if the collection service handles the form itself.
Whichever option you pick, write down four things. First, who owns open requests and who can access the service. Second, the list of payees who have been sent a request, so gaps are visible at a glance. Third, where each status change gets recorded, which should be the payee's file rather than a chat thread or a personal note. Fourth, what happens when a returned form looks wrong.
A workable sequence for a single request looks like this. Send the request and log it with the payee, the date sent, and the person responsible. Then either check or wait for the change, depending on your option. When the status shows the form has come back, review it against the instructions your team uses and record the outcome in the file. If nothing has come back by the time your review runs, the request goes onto a follow-up list with a named person, and that person decides how and when to reach the payee again. At every stage, the request should have a visible next owner. If you cannot point to one, the process has a gap.
Hand-offs deserve their own attention. Staff change roles, people go on leave, and a request that one person was tracking can fall out of view. Keep the open list and the payee files in a place the whole team can see, rather than in one person's mailbox or spreadsheet. A new owner should be able to pick up any request by reading the file, without asking the previous owner what they meant.
When a returned form looks wrong
Sometimes the form comes back and something on it does not look right. The IRS instructions give one specific rule that can guide the team's thinking, though it covers only one line. For line 3b, they say, "You may rely on the information provided on line 3b unless you know that it is incorrect" ( Instructions for the Requester of Form W-9).
Some returned forms are not wrong so much as incomplete. A missing line or an unclear answer is a different problem from a value that contradicts your records, and it usually calls for a narrower request back to the payee. Sorting the two kinds of problem at intake keeps reviewers from treating every gap as a dispute.
That rule applies to line 3b. Your team should decide in advance who reviews these cases, what they check the form against, and who contacts the payee. Writing that down prevents a situation where each person guesses differently and the payee hears from the firm in inconsistent ways. Keep the reviewer's decision in the payee file too, so the reason for accepting or questioning a form is visible later. For related review steps, see vendor W-9 review.
When you write to the payee, keep the message short and specific. Say which form you are asking about, what needs to be corrected or completed, and where to send it. Avoid repeating sensitive numbers in the message itself, and point the payee to the form and the return method your team uses instead.