An import task that finds fewer listings than you expect, or none at all, is almost always the filters rather than a fault. Every filter you set is sent to the MLS as part of the query, and the MLS returns only listings that satisfy all of them at once.
Work down the causes below in order. The first three account for nearly every case.
Before you start
- Both connection banners on MLS Import Settings are green. A red MLS banner is a different problem.
- You have one real example: an MLS number that exists in the MLS and that you expect this task to import.
1. A filter value does not match the MLS wording
Filters are exact matches, not searches. Miami matches listings whose city is stored as Miami and nothing else. miami, Miami Beach and MIAMI are different values as far as the query is concerned.
Two rules govern how filters combine. Several values in one field are treated as any of these, so a city list of Miami and Coral Gables returns both. Different fields are treated as all of these, so a city of Miami with a property type of Residential Lease returns only listings that are both.
Take the listing you know should import, open it in your MLS, and read the values it actually holds for each field you filtered on. One mismatch anywhere in the set removes it from the results.
Diagnose by subtraction. Clear every filter except one, run the task, then add the others back one at a time. The filter that takes the count to zero is the one whose value does not match your data, and you find it in three runs instead of guessing.
2. The internet display filters are set to yes
Internet Entire Listing Display and Internet Address display are the two fields sellers use to opt a listing out of internet publication. They behave differently from every other filter on the task.
- Set to yes, the filter is added to the query and only listings the MLS marks as explicitly allowed come back.
- Set to no, no filter is added at all. The field stops being part of the query and the MLS decides what to return under its own rules.
That asymmetry is what produces the empty result. If your MLS leaves the field blank rather than setting it to true, asking for the allowed ones excludes every listing, including the ones that are genuinely publishable. Set both to no and run the task again. If listings appear, this was the cause.
One provider is treated differently. On an RMLS feed the entire listing filter is never sent, so setting it makes no difference there.
3. The listing’s status is not one you selected
Status is a filter like any other, and MLS status wording varies. A task set to Active alone will not import a listing the MLS has moved to Active Under Contract, even though most people still think of it as on the market.
Open the listing in your MLS, read its exact status, and confirm that value is selected on the task.
4. The listing does not reach the API
MLSImport reads your feed through the MLS API and has no other route to the data. If your MLS excludes a listing before it answers, that listing cannot be imported and nothing on your site can change it.
Boards apply their own rules on top of the fields you can see, which is why a listing can be visible to you inside the MLS system and absent from the API. Nothing in the response says why. When the filters check out and a specific listing still never arrives, that question goes to your MLS or your API provider with the MLS number in hand.
5. You are looking at the hourly sync rather than a full run
The hourly sync asks only for listings changed since the last run, which is what keeps it fast. A filter you widened this morning does not pull in the older listings it now matches until the task runs a full import.
After changing filters, run the task manually once and let it finish. From then on the sync keeps it current.
A run that returns nothing changes nothing. It does not delete properties, and it does not mark anything as gone from the MLS. Removal is decided by the daily reconciliation pass against the full list of MLS identifiers, never by what a single import run happened to return.
If it does not work
- The count is right but properties are missing from the site. Another task claimed them first and they are on the site under a different task. Search by MLS number in the properties list before assuming they were never imported.
- The task imports some cities and not others. The missing city is spelled differently in the feed, often with an abbreviation or a saint prefix written out.
- Everything imports except new listings. The task is filtered on a listing ID or a narrow date driven field left over from testing.
- Nothing imports and the run ends immediately. Check the field selection has been saved at least once. A task cannot write properties before the field mapping exists.
Frequently asked questions
Can you tell me why my MLS excluded a particular listing?
No. The API answers with the listings that match and gives no reason for the ones it leaves out, so a missing value and a deliberate exclusion look identical from our side. Your MLS or API provider can see both.
If I narrow a filter, are the properties it no longer matches removed?
Not by the import itself. They stop being updated by that task, and the daily reconciliation decides their fate from your status settings rather than from the filter change. Use the delete tool if you want them gone straight away.
Why does the same listing import for one task and not another?
Because each task sends its own filter set. A listing has to satisfy every filter on the task that is running, so a broad task can hold listings a narrow one will never see.
Does a large filter set slow the import down?
No. Filters are applied by the MLS before it answers, so a tighter filter usually means less data and a faster run. What slows an import is the number of listings returned and their photos.
Related
- Controlling which listings are imported
- Avoiding overlapping import tasks
- Closed and off-market listings
- How daily reconciliation works
- Import History and the activity log
Send us the import task and one MLS number you expect it to import and we will read the query it builds.