Skip to content
All Guides And Research

Product & Technology

AI Manifest Analysis: What To Check Before You Trust The Total

Before you trust an AI manifest total, trace one row: the seller file, the product identity, the evidence behind its value, the coverage of the total, and the costs only you can supply.

Liquidata Team Published Updated 7 min read

Every liquidation buyer knows the routine. A new auction goes live, you download the manifest, and you open it in a spreadsheet. Then you start copying identifiers into search bars one at a time, checking what items actually sell for, adjusting for condition, and trying to decide whether the load is worth your bid. Many manifests make it harder: poorly formatted PDFs, inconsistent column headers, missing UPCs and descriptions like “electronics accessory.”

Software can do the lookup work across every row. That makes the result faster to reach, not automatically right. An analysis total is only as good as the rows behind it, and the quickest way to judge a total is to trace one of those rows from the seller’s file to your bid.

The audit below follows one row from a published Liquidata analysis through each check.

One Row, Five Checks

Row 32 of the Small Appliances Customer Returns Liquidation Pallet — 42 Items analysis, observed .

  1. 01File

    What Did The Seller’s File Say?

    Row
    32 of 42
    Product
    Wyze 24Kpa Powerful Suction Cordless Vacuum Cleaner, Lightweight Stick with HEPA Filter
    Qty
    1
    Condition
    customer returns
    Seller Price
    $139.00

    Check. The analysis should keep the seller’s row as written, beside anything it adds. If you cannot find the original text, you cannot check the result against it.

    Still Open. The seller’s text is a claim. It does not yet identify a product.

  2. 02Identity

    Which Product Is It?

    Model Field
    24Kpa Powerful Suction Cordless Vacuum Cleaner

    Check. An exact identifier supports an exact match. This model field holds a description instead of a model number, so it can support, at best, a close match.

    Still Open. Which exact product this is.

  3. 03Evidence

    What Supports The Value?

    Evidence
    Approximate
    Fair Price
    $46.00
    Seller Price
    $139.00

    Check. Approximate means Liquidata estimated the value from a comparable or derived price because exact product evidence was incomplete. Here the researched price is 66.91% below the seller’s price. A gap that large is a reason to look at the row, not to trust either number.

    Still Open. Whether the comparable product is close enough to this one.

  4. 04Coverage

    How Much Of The Total Is Supported?

    Fair Value
    $5,576.59
    Rows In It
    41 of 42
    Evidence Mix
    36 Verifiable · 4 Approximate · 2 Unverifiable

    Check. Read the total with its coverage. Fair Value adds the researched prices on 41 of 42 rows, and this Approximate row is one of them. Row 19 has no researched value, so it is not in the total and is not counted as $0.

    Still Open. What the row without a researched value is worth. You would pay for it: the seller’s value on it is $90.

  5. 05Your Costs

    What Will It Cost You To Land And Sell?

    Ask
    $1,600 for all 42 rows

    Check. Fair Value is researched retail evidence, not what you will sell for. Subtract freight, buyer fees, labor, storage, selling fees and your target profit from what you expect to recover. The bid ceiling worksheet does the subtraction with your numbers.

    Still Open. Your freight quote, your channel’s fees, and the price a returned unit brings in its actual condition. Only you can supply them.

Why Manual Research Is Hard To Scale

Researching a manifest by hand is slow, and the work gets worse as the file grows.

Fatigue errors. Late in a long file, it is easy to rush the last rows, spot-check a category you know less well, or miss a product that needs a closer look.

Missed opportunities. Time spent on one manifest is time not spent on the next one. The speed of your research limits how many loads you can evaluate.

Stale prices. Resale prices move. The sold listings you checked last week may not reflect this week’s market.

These are the reasons buyers turn to analysis tools. They are also the reasons to audit what a tool gives back.

Five Questions To Ask Of Any Total

1. Which rows are in it? A total built from 41 of 42 rows is a different claim from one built from 20 of 42. Look for the coverage figure beside the total, and for the rows left out.

2. What does the number measure? The seller’s value, a researched retail value, an expected resale estimate and the asking price measure different things. A tool that blends them into one figure has hidden the question you need to answer.

3. How strong is each row’s evidence? An exact identifier match is stronger evidence than a match on a vague description. The analysis should say which kind each row has.

4. What happened to rows without evidence? A row the tool could not research should show no researched value. Zero is a real amount; an empty value means unknown. If a tool fills the gap with a guess or an average, the total includes value nobody checked.

5. Whose costs are in it? Freight, labor, storage, selling fees and your target profit belong to you. A bid figure that does not ask for them is not your bid.

What Software Can Catch That A Rushed Review Misses

Inflated values. Manifests often list a retail value well above what a product sells for. Researching every row against current market evidence shows where the seller’s value and the researched value disagree, and by how much.

Recalled products. No one wants to check a 200-row file against recall notices by hand. Liquidata’s CPSC Recall Cross Check screens every analyzed line item against the available CPSC recall records, keeps an identified product match separate from a recent-brand warning, and flags rows without enough identity instead of clearing them. No match is not a safety certification; Check Recalled Products Before Resale shows how to compare one product with an official notice.

Category concentration. A file that looks varied can be dominated by one low-margin category once every row is classified. Read the category mix before the headline total.

What Liquidata Does With A Manifest

Liquidata Manifest Analysis accepts an Excel, CSV or PDF seller manifest and maps the seller’s fields without requiring you to rebuild the product list by hand. It checks every line item against market evidence and uses the supported matches to determine the load’s Fair Value. The original seller row stays attached to each result.

  • Evidence classes. Each row is Verifiable, Approximate or Unverifiable, so you can see how each value was supported.
  • Coverage and confidence. The share of the file that resolved is reported beside the total. Unsupported rows can keep a null researched price; Liquidata does not invent a comparable value to fill the row.
  • Separate measures. Seller Value, Fair Value, Expected Resale Revenue, acquisition cost and the Bid Recommendation keep distinct meanings. Expected Resale Revenue is a screening estimate, not guaranteed revenue or profit.
  • Your costs. The Bid Recommendation uses the desired profit, transport, storage and fulfillment inputs you enter. If those inputs leave no profitable case, it says “No profitable bid.”
  • Portable output. You can export the analyzed rows to Excel.

For loads without a manifest, Manifest Generation turns images or PDFs into candidate rows that you review, with confidence and source references, before you finalize a manifest.

After You Buy

The most useful check on any analysis is what happened next. Did the load recover what you projected? Which rows arrived short, damaged or different from the description? Compare your projection with your actual results load by load, and your next projection gets better.

Liquidata Purchase & Vendor Records keeps each purchase connected to the analysis it came from, with its vendor, documents, notes, reminders and the status you maintain. The comparison itself is yours to make; the record keeps the evidence together while you make it.

Is It Worth It For Your Business?

If you buy one or two loads a month, careful manual research may be enough. You have time to check the rows that matter.

If you buy several loads a month, the research time adds up, and an analysis that covers every row lets you spend your attention on the exceptions: unresolved rows, approximate matches and large gaps between the seller’s value and the researched one.

If you run a bin store, buy truckloads or manage a team, a consistent row-level record matters as much as speed. Everyone reviewing a load should be looking at the same evidence, the same coverage and the same costs.

In every case, the audit questions above apply. A faster total is useful only when you can see what it is made of.

Audit Your Own Manifest

Upload a seller manifest and trace any row the same way: the original text, the matched product, its evidence class, the coverage of the total, and the costs you enter.

Analyze Manifest
← All Guides And Research