Indiegogo InDemand Bag Fulfilment: Late Backers and Re-orders

InDemand turns fulfilment into a pipeline

Most bag fulfilment advice assumes an event. A campaign closes, one order is placed, one inspection happens, one dispatch goes out, and the project ends. InDemand does not work that way. After the funding period closes, the campaign page can continue taking orders, which means the demand signal does not stop at the close date and the fulfilment operation has to keep answering it. The result is a pipeline with three distinct flows: the initial backer dispatch, the continuing InDemand orders, and the periodic re-orders that keep stock available for both.

The change sounds administrative and it is actually structural, because it alters what a production order is for. In the event model, the production order exists to satisfy a known list of backers. In the pipeline model, the production order exists to satisfy a known list plus an unknown quantity, and the planning question becomes how much of the unknown to cover with the first run. Getting that wrong in either direction has a cost: covering too little means the continuing orders arrive to an empty shelf and a 35 to 50 day gap, and covering too much means capital sits in cartons that the demand rate may not justify.

The correct response is not to guess the continuing demand more accurately. It is to design the programme so the guess matters less, which means three specific things: a reserve held back from the first run, a re-order trigger defined in advance rather than decided each month, and a reference retained between runs so that the second production order reproduces the first. The rest of this guide works through those three and the details that surround them. For the platform context and how InDemand fits alongside a second platform, the Indiegogo bag campaign guide is the starting point.

Late backers: what they create

A late backer is not simply a backer who arrived after the close date. In an InDemand programme they are the demand signal itself, and their behaviour differs from launch-period backers in three ways that matter operationally. They arrive at a lower and steadier rate rather than in the bursts a campaign generates. They convert on the finished product and its reviews rather than on the promise, so their tier mix tends to shift towards whatever was best received. And they expect shorter delivery, because from their perspective the product already exists.

That last expectation is the one that causes the trouble. A late backer sees a product that has been photographed by hundreds of earlier backers, reads a delivery estimate in weeks, and reasonably assumes the goods are sitting on a shelf. If they are not, the gap is 35 to 50 days for bulk plus freight, and the support conversation that follows is more expensive than the reserve stock that would have prevented it. The fix is not to hide the lead time. It is to state it in the listing and to hold enough reserve that the stated lead time is usually beaten rather than usually missed.

• Lower arrival rate. A steady trickle rather than launch bursts, which favours consolidated re-orders over frequent small ones.
• Shifted tier mix. Late buyers pick what reviewed well, so the colour and size split will not match the campaign's original split.
• Shorter expectation. They believe the product exists and expect dispatch in weeks, which makes reserve stock a customer service measure.
• Higher return sensitivity. Late buyers have read the reviews, including the complaints, and will check the specific things that were criticised.

The shifted tier mix deserves a specific note, because it is where a re-order goes wrong. If the campaign sold 40 percent of units in one colour and the reviews then praised a different colour, the continuing demand will not follow the original split, and a re-order placed on the original proportions will produce the wrong stock. Track the continuing orders by variant from the first month, and reorder on observed proportions rather than original ones. The bag MOQ calculator is a useful way to run that split against the minimum before committing.

Re-order maths: the minimum applies again

The minimum does not expire. 500 pcs per style and per colourway applies to every production order including the fourth one, and this is the constraint that shapes an InDemand programme. A continuing demand of 40 units a month accumulates to 480 units a year, which sounds like a healthy business and is, in production terms, one order a year placed just under the minimum. The arithmetic works only if the re-order is consolidated and timed, rather than placed in response to each month's orders. A sample for a new variant, by contrast, takes 6 to 10 working days, which is the cheap part of the calendar and the part worth doing early.

The last row is the one most InDemand programmes overlook and it is usually the cheapest. If the campaign has a second production requirement, a colourway extension, a new format, or a top-up for a second platform, adding 500 units of the continuing SKU to that open production slot costs less than opening a slot for it alone, because the pattern, the tooling and the setup are already being paid for. The practical habit is to tell your programme coordinator what the continuing SKUs are, so that a slot opening anywhere in the programme can carry them.

The other side of the arithmetic is what a re-order costs in time, and the number to plan with is 35 to 50 days for bulk from release plus freight. A programme that re-orders when stock reaches zero has a 35 to 50 day stockout plus transit. A programme that re-orders at a defined trigger, say when reserve falls to 25 percent of the last re-order quantity, absorbs the lead time inside the remaining stock. The trigger, not the lead time, is the thing under your control, and setting it is a five-minute decision made once.

Holding a golden reference between runs

A re-order reproduces whatever the production team is given. If they are given the specification and a retained physical reference, the second run matches the first. If they are given the specification and a photograph, or the specification alone, the second run approximates the first, and the approximation usually shows up in the things a specification cannot fully express: the hand of the fabric after washing, the tension of a binding, the exact shade under retail lighting, the way a strap lies when the bag is empty.

The retained reference is usually called the golden sample, and there are three rules for keeping one across an InDemand programme. Keep two, not one, and store them in different places, because a reference that is used for photography and comparison will wear. Keep them in a dark, dry place, because the most common failure of a retained reference is that it has faded or yellowed and is then used as the standard it no longer represents. And record what it is: which production lot it came from, which specification version it was approved against, and the date it was approved.

The reference also has to be current. If a change was agreed during the first run, the reference has to be updated in the same week or it becomes the record of a product that is no longer being made. On a pipeline programme this is the point at which most drift enters, because the change happens during a busy dispatch period and the reference update is deferred, and 6 months later the second run is built to a sample that predates an improvement. The sample approval reference sets out what to check on a retained unit before it is used as a standard.

Batch drift between run one and run two

Even with a perfect reference, two runs made months apart differ, and the differences cluster in four predictable places. Fabric comes from a different lot and the hand changes slightly. Thread and webbing come from different batches and the shade changes slightly. The operator mix on the line is different, so the stitch tension and the binding width vary. And trim, particularly zips and buckles, may be sourced from an alternate supplier if the original was unavailable, which changes the pull feel and the colour match even when the specification was followed.

None of these are defects, and that is precisely the problem. Each is individually within tolerance and collectively visible, particularly when a late backer receives a bag and compares it with a photograph an earlier backer posted. The control is not to demand zero variation, which is not achievable and would be expensive to attempt. It is to inspect the second run against the retained reference on the four things that drift, and to record the comparison, so that a decision about whether the variation is acceptable is made deliberately rather than by default.

The inspection standard itself should not change between runs. AQL 2.5 to ISO 2859-1 applies to the re-order exactly as it applied to the original, and the defect register should continue from the first run into the second. A re-order is where inspection discipline most often relaxes, because the quantity is the same, the product is familiar and everyone involved wants to ship. The consequence of relaxing it is that the second shipment, which goes to the buyers with the highest expectations, receives the weaker inspection. For the sampling detail, see the AQL reference for bag campaigns.

Colour continuity across dye lots

Colour is the drift that backers notice first and the one most often caused by a decision nobody thought was a decision. A fabric ordered at 500 units is one dye lot. A re-order at 500 units six months later is a different dye lot of the same nominal colour, and the difference between two lots of the same reference is normal in the industry and visible in a photograph. The tolerance is usually expressed as a grey scale rating, and the practical planning figure for an acceptable lot-to-lot difference is a rating of 4 or better on the grey scale for colour change.

There are three ways to manage it and they have different costs. The cheapest is to accept lot-to-lot variation as normal, and to say so plainly if a backer asks, which is an accurate and reasonable answer. The second is to order both the original and the re-order fabric in one purchase, so both runs come from the same lot; this ties up fabric and storage, and it only works if the re-order quantity was known at the time, which in an InDemand programme it usually was not. The third is to specify the tolerance in the material call-out and require a shade band approval before the re-order fabric is cut.

The third is the one we would recommend for a continuing programme with a visible brand colour. A shade band is a small physical card cut from the approved lot and attached to the specification, and approval means the new lot falls within the band under a specified light source. It costs a few days if the lot is outside the band, and it prevents the situation where 500 finished bags in the wrong shade are the first evidence that the lot had moved. Where the brand colour is a print rather than a fabric, the same discipline applies to the print, which is covered in the print colour fastness guide.

Reserve stock versus just-in-time

The reserve decision is the one that most directly determines whether an InDemand programme feels reliable to its buyers. Holding no reserve means every continuing order is either served immediately from chance leftover stock or enters a 35 to 50 day queue. Holding a small reserve means most continuing orders dispatch immediately and the re-order can be timed to the trigger rather than to the stockout. Holding a large reserve means capital is tied up and, in a bag programme, exposed to the risk that the next revision or a change in taste makes it slow-moving.

The planning figure to start from is 5 to 10 percent of the annual continuing volume, held as finished goods, plus whatever is needed to cover the lead time at the observed demand rate. At 40 units a month and a 50 day lead time, the lead-time cover is about 67 units, so a reserve in the range of 60 to 110 units is a reasonable target. The two figures to combine are therefore the lead-time cover, which is arithmetic, and the variance cover, which is a judgement about how bursty the demand turns out to be.

Reserve also serves three purposes that have nothing to do with continuing demand, and they should be counted when sizing it. Replacements for units damaged in transit, which in most bag programmes run at a low single-digit percentage. Address failures and returns that come back and need to be re-dispatched or resold. And the inspection shortfall, which is the units that fail AQL and reduce the shippable quantity below the ordered quantity. A reserve sized only for continuing demand will be consumed by these three before it ever serves a late backer.

Freight modes for a continuing pipeline

The event model usually resolves freight once, at the point the bulk goods are ready, and the decision is essentially sea or air on one shipment. A pipeline resolves it repeatedly on smaller quantities, and the economics change. A re-order of 500 units is a small consignment in freight terms, and the fixed costs of a shipment are spread over a third of the quantity, so the per-unit freight cost of a continuing pipeline is structurally higher than the per-unit cost of the original dispatch. That is a cost of the model and it belongs in the continuing margin, not a surprise at the freight invoice.

The practical response is to match the mode to the urgency of each flow. The original dispatch, which has a published date, usually goes by sea with air held as contingency for a shortfall. A re-order that is being placed because reserve has fallen to the trigger is not urgent in the same way, because the reserve is covering the demand, so it can also go by sea and the freight saving is real. Air becomes correct when the reserve has already been consumed, and at that point the cost is the price of a promise that was made by holding too little.

The third freight consideration is consolidation. If a programme has several SKUs continuing through InDemand, consolidating their re-orders into one shipment reduces the fixed cost per unit considerably, and it also simplifies customs clearance, which is a fixed administrative cost per entry regardless of size. The customs and duties reference for imported reward bags covers the entry side. The consolidation discipline is the same as the production one: one trigger date, one shipment, several SKUs.

Payment terms on repeating orders

The standard terms are T/T with 30 percent deposit before bulk begins and 70 percent balance before shipment, and they apply to a re-order in the same shape as they applied to the original. What changes on a continuing programme is the frequency and therefore the cash behaviour. A programme placing two re-orders a year has two deposit dates and two balance dates, and the cash requirement is cyclical rather than one-off. Modelling it on a rolling basis, rather than treating each re-order as a separate event, is what prevents the second deposit arriving before the revenue from the first has landed.

There are two things worth asking about on a repeating programme. The first is whether a deposit can be held on account for a scheduled re-order, so that a slot can be opened without a separate payment round each time. This is not always offered and it depends on the relationship and the history, but where it is available it removes several days from each cycle. The second is FOB Xiamen as the trade term, which fixes where the cost and the risk transfer, and which should be stated identically on the original order and every re-order so the freight comparison between them is meaningful.

The third point is about the quotation itself. A re-order quotation should be requested against the original specification version, and it should be compared to the original quotation line by line rather than as a single number. If the unit price has moved, the reason is usually one of three things: the fabric lot price moved, the quantity is different, or the trim was sourced from an alternate. Each has a different implication, and only the line-by-line comparison tells you which. A single number that has gone up by 4 percent is not information. It is a question.

When to close InDemand and consolidate

There is a point at which continuing InDemand demand stops justifying a production pipeline, and it is usually reached before the creator recognises it. The signal is not a fall in the order rate alone, because a fall can be seasonal. It is the combination of three things: the observed monthly rate falling below the rate needed to clear 500 units in the holding period you can tolerate, the reserve taking longer to deplete than the re-order lead time, and the variant spread widening so that each additional colour becomes its own minimum problem.

At that point there are three honest options. Continue, but consolidate the range to the one or two variants that still move, which reduces the number of minimums the programme supports. Move the continuing sales to a direct channel where a lower volume is sustainable, which changes the fulfilment operation but not the product. Or close InDemand, fulfill the remaining orders from a final consolidated re-order, and treat the product as a finished campaign. None of these is a failure. The failure is continuing to hold a pipeline open for a demand rate that cannot support the minimum, because the cost does not appear as a line item and so it goes unexamined.

The consolidation decision should also account for what the product is worth later. A bag that fits a common laptop size or a standard device form factor will keep selling outside the campaign and supports a continuing pipeline. A bag tied to a specific campaign revision, a specific device or a specific moment does not, and the right move with it is a final re-order sized to the remaining demand plus a small reserve, and then a clean close. On a two-platform programme the same logic applies per platform, and the cross-platform sampling guide covers how the file is kept consistent across both.

An InDemand operating calendar

The calendar below is organised by trigger rather than by date, because a continuing programme has no fixed end and a calendar of dates becomes wrong the first month a date moves. Each stage fires when its condition is met, and the conditions are all measurable from the order log you are keeping anyway.

1. At first dispatch. Hold 5 to 10 percent as reserve, recorded separately, and retain two physical golden references from the approved lot.
2. From month one. Log continuing orders weekly by variant, not monthly in total, so the re-order split reflects observed demand.
3. When reserve reaches 25 percent of the last order. Request the re-order quotation against the original specification version, line by line.
4. On quotation acceptance. Release bulk and start the 35 to 50 day window. Confirm FOB Xiamen and the T/T 30/70 schedule in writing.
5. Before cutting. Approve the fabric lot against the retained shade band and approve the trim against the retained reference.
6. At inspection. Run AQL 2.5 to ISO 2859-1, continue the same defect register, and record the comparison against the golden reference.
7. At receipt. Top the reserve back up before serving continuing orders, not after.
8. Quarterly. Review the rate, the reserve depletion and the variant spread, and decide whether to continue, consolidate or close.

The two lines in that sequence that do most of the work are the weekly variant log and the reserve top-up before serving continuing orders. The log makes the re-order split correct instead of inherited, and the top-up order makes the reserve a genuine buffer rather than a number that is calculated once and then quietly consumed. Everything else in an InDemand programme is ordinary production work. Those two are the parts that are specific to a pipeline and the parts most often skipped.

Frequently Asked Questions

People Also Ask