Features
Part of Campaign management tools: what they take over, step by step
How influencer agencies can test campaign management tools
How influencer agencies test campaign management tools: script a failed campaign, add disclosure and usage-rights checks, compare finalists, set a deadline.
There is no ranked list on this page, and the reason is not modesty. A campaign management tool is bought to fix one specific bottleneck in one specific agency, and the same product that saves a team of twenty an afternoon a week will slow a team of three down.
A ranking would have to average across buyers whose problems have nothing in common. What can be written down is the method: how to build a test the product can genuinely fail, and what to look at once two candidates have both passed it.
What to take away
- The test has to be written before you see a demo. After the demo you will write a test the demo passes.
- Run the evaluation on the campaign that went worst last year, not on a clean one.
- Two finalists are separated by what happens when something goes wrong, and that never appears in a feature list.
Why no ranking is offered
Three things would have to be true for a ranked list to help you, and none of them are.
The buyers would have to want the same thing. An agency whose pain is chasing drafts and an agency whose pain is paying forty people in six currencies are shopping in different markets that happen to share a category name.
The ranking would have to survive contact with your data. Most of what separates these products is how they behave with your volume, your client mix and your existing spreadsheets, and none of that is visible from outside.
The list would have to be maintained at the speed the products change. Any page that claims a current order is either being rewritten constantly or is stale, and you cannot tell which from reading it.
So this page gives you the method. The general version of that method, including how to compare things that are not the same kind of thing, is set out in the guide to structured vendor comparison.
Build the test from your worst campaign
Pick the campaign that went worst in the last year and write down what actually happened, in order. That document is your test script. It contains the failures a demo is designed to avoid.
Test script from worst campaign
- Brief changed twicefind version one again
- Creator stopped replyingdoes silence have a state
- Draft four revisionswho approved which version
- Verbal rate contradicts contractdoes system flag
- Late paymentwas delay visible beforehand
- Client asked everything backexport full record
| Stage of that campaign | What to make the product do | What you are watching for |
|---|---|---|
| The brief that changed twice | Load version one, then change it, then find version one again | Whether the change is visible to everyone or only to whoever made it |
| The creator who stopped replying | Enter a person who is neither active nor rejected | Whether silence has a state, or only a blank field |
| The draft that went through four revisions | Attach four versions with comments from two people | Whether you can tell which version was approved, and by whom |
| The rate that was agreed verbally | Record a fee that contradicts the one in the contract | Whether the system flags the conflict or stores both quietly |
| The payment that went out late | Run a payout, including the tax paperwork | Whether the delay would have been visible before it happened |
| The client who asked for everything back | Export the whole campaign, messages included | Whether what comes out is a record or a screenshot of one view |
Write the pass condition for each row before the first call. This is acceptance testing applied to a purchase: the criteria are fixed in advance so the result is a finding rather than an impression.
Influencer-specific rows to add to the script
The six rows above would test almost any workflow product. A campaign management tool for an influencer agency also has to survive the parts of the job that only creator work has, so add these rows before the first demo.
- Disclosure checks.The Federal Trade Commission's staff brochure Disclosures 101 for Social Media Influencers, which sits alongside the FTC's Endorsement Guides, says an endorsement should make it obvious when the creator has a relationship with the brand, and that a creator should disclose the relationship if they got anything of value to mention a product, including free or discounted products. Make the product record a gifted-product campaign and watch whether the disclosure requirement follows the creator into the brief and the approval step, or has to be remembered.
- Disclosure placement.The same guidance warns creators not to mix a disclosure into a group of hashtags or links. Load a draft caption whose disclosure sits at the end of a hashtag block and see whether the review step can flag it or only stores it.
- Account permissions.Connect a test creator's social account, then revoke the permission halfway through the campaign. Watch whether the product tells you that reporting has stopped or keeps showing the last numbers it had.
- Usage rights.Record content licensed to the client for a fixed period, then move the date past the end of that period. A pass is a warning before the rights lapse, not a note in a free-text field.
- Creator data.Ask who can see a creator's personal details, rates and payment information, and remove one creator entirely. Check that the removal reaches the exports as well as the screens.
These rows are where products built for brand marketing teams and products built for agencies tend to part company.
Four criteria that separate finalists
Once the obviously wrong products are gone, feature lists stop distinguishing anything. Four things still do.
Four finalist criteria
Criterion
- Missing facts
- blank, guess or unknown
- Concurrent action
- two people at once
- Process insistence
- match or expensive mismatch
- Leaving
- export during evaluation
What to test
- Missing facts
- Concurrent action
- Process insistence
- Leaving
- Where the product puts a fact it does not have. A blank, a guess and an explicit unknown produce very different behavior six months later.
- What it does when two people act at once. Agency work is concurrent and most single-brand products assume it is not.
- How much of your process it insists on. A product with strong opinions is fine if they match yours and expensive if they do not.
- What leaving looks like, tested rather than described. Ask for the export during the evaluation, not during the notice period.
The depth behind a vendor's coverage claims is a separate question with its own traps, worked through in the guide to platform coverage and permissions.
Running the evaluation without losing a month
Set the end date before you start. A comparison with no deadline turns into a research project, and the longer it runs the harder it gets to walk away from the option you have already invested in looking at. That is escalation of commitment, and it decides more purchases than any scoring matrix.
Evaluation rules that save a month
- Set end date before starting
- Name one decision-maker, not a committee
- Same test script and data for all
- Drive the product in one session
- Ask reference from someone who left
- Put quotes on a single basis
- Name one person who decides. A committee produces a compromise product nobody wanted.
- Give each candidate the same test script and the same data. Different demos are not comparable evidence.
- Insist on doing the driving in at least one session. Watching a vendor use their own product tells you what it looks like in expert hands.
- Ask for a reference from someone who left. The vendor will not enjoy the question and the answer is worth the discomfort.
- Put the quote on a single basis before you compare two of them, such as the cost per campaign at your real volume, whether the vendor charges per seat, per creator, per campaign or by managed spend, because the meters differ more than the prices do. The way those meters are built is set out in the guide to how these tools are priced.
What a demo cannot show you
A demo shows the product in its best month with its cleanest data. What decides whether you still like it in month three stays invisible in that room.
Three things decide:
- How support answers when you are wrong rather than when the product is broken.
- How it behaves at the end of a quarter when everyone is in it at once.
- How much of your team's day is spent maintaining the tool rather than using it.
That last one is the quiet cost, and every system of record needs somebody to keep it true. If nobody is named, records rot, and a rotting record is worse than a spreadsheet because people still trust it.
Deciding who owns the system after launch belongs in the purchase, not after it. Name that owner, and the hours a week they will spend keeping records true, before you sign.
Common questions
Can we just pick whatever the biggest agencies use?
Their constraint is coordination across hundreds of people. Yours probably is not. Copying the choice imports a product built for a problem you do not have, at a price set for a budget you do not have either.
How many products should we look at?
Enough to learn the shape of the market, then two. Beyond two, you are comparing preferences rather than evidence, and the extra weeks cost more than the difference between the finalists.
The vendor will not run our test script. Is that a red flag?
Not always. It is a fair signal when the refusal comes with no alternative. Ask them to propose one instead, and judge the proposal.
What if none of them fixes the actual bottleneck?
Then the answer is not a purchase. Some bottlenecks are staffing or process problems wearing a software costume, and buying a tool for them adds a subscription to a problem you still have.
