Small media budget: under €10k a month, the problem is not the amount, it is the spread. What 2026 really allows in social, search, TV and DOOH, with data.
"We have €8,000 a month. Is it even worth doing media?" It is the question I hear most often in a first meeting, and it is the wrong question. €8,000 a month is close to €100,000 a year. That is not a small budget. It is a budget that does not forgive spreading. And spreading is exactly what I see in nine accounts out of ten at that level: Meta, Google, LinkedIn, a bit of TikTok, a trade show, two freelancers. Every line at €1,500. None of them ever rises above the noise. The problem is not the amount. It is the number of lines.
Small media budget: the real ceiling is the learning phase
Start with a rule everyone knows and almost nobody applies to their own budget. Meta states in its help centre that an ad set needs roughly 50 optimisation events within 7 days to exit the learning phase. Below that, it stays in "learning limited", which means the algorithm bids blind. Now run the numbers with a real cost per lead. The Talyco observatory (published 7 July 2026, €460,000 of real spend analysed over twelve months across French advertisers) gives a median cost per lead of €23.80 on Search and €8.30 on Performance Max, with a range from €3 to €213 depending on the business. Take €25 per lead. 50 leads a week is €1,250 a week, about €5,400 a month. For a single ad set.
Under €10,000, you can therefore afford to let one ad set learn, two at most. Not ten. The account I see most often is ten ad sets at €800 each, all in "learning limited" for six months, and a founder concluding that "Meta doesn't work for us". Meta does not work at €800 per line. It never got the chance to try.
What €10,000 really buys in 2026
The Talyco medians give the scale. Meta CPM: €3.40. Facebook CPC: €0.12, Instagram: €0.28. Search CPC: €0.58. At those prices, €10,000 on Meta is close to three million impressions. On Search, about 17,000 clicks and, at the median CPL, four hundred leads. Those are not small-budget numbers. They become small when you cut them into ten.
And offline is no longer closed at this level. M6 Unlimited advertises a TV campaign "from €5,000". On 19 June 2026, TF1 Pub launched with the CPME du Rhône a pilot of self-service addressable TV and streaming, TF1 AdManager, with a €1,000 entry ticket and geo-targeting on the business's catchment area. FranceTV Publicité sells local campaigns via ADspace from €2,000, and 11.4 million set-top boxes were eligible for ad substitution in the first half of 2025 (Radiopub, 9 September 2026). On screens, Q1 2026 as measured by ACPM counts 50,000 connected DOOH screens, 306 campaigns, and 12% new advertisers. On our own rate card, a single screen starts around €2,000 a week, at a €4 to €9 CPM.
But "possible" does not mean "useful". A €1,000 TV test measures nothing: too few contacts, too little repetition, no readable signal in your sales. If you go into TV or DOOH under €10,000 a month, you go in bursts: a full month on a single area, then nothing. Not €1,000 a month for ten months.
Geography is the only free multiplier
There is one lever that costs nothing and changes everything at this budget level: the area. €10,000 across France is noise. €10,000 on a metro area of 200,000 people is a presence. For Comptoirs de la Bio, isochrone targeting around each store halved the budget compared with printed flyers, with 100% tracked delivery. The flyer sprayed. The isochrone aims. The same reasoning holds for a scale-up without stores: where your customers are, where your delivery is good, where you already have references. For Deblock, we did the opposite of a national campaign, encircling metro stations with 100-metre mobile geo-targeting, more than 30 million impressions concentrated where the target is dense. That was a big budget. The principle scales down without losing its shape: density before reach.
One channel, one area, one message: the 70/20/10
Here is the split I defend under €10,000. 70% on a master channel, the one where your target spends time and where you can exit the learning phase: most often geo-targeted social, sometimes Search when demand already exists. 20% on the channel that catches what the first one triggered: brand search and category queries on the area. 10% for testing, but set aside over a quarter to fund a real burst, not spent in crumbs every month.
And at this level, creative is the only lever you fully control. For Essity, a gamified social mechanic generated 14 million impressions at about €6 per lead. That price does not come from the auction, it comes from the mechanic and the creative. Under €10,000, three to five strong creatives tested in one campaign, not thirty variants across ten ad sets.
The €8,000-a-month plan I would sign
€5,600 on Meta, one campaign, two ad sets at most, four creatives, one isochrone area. €1,600 on Search for brand and category, same area. €800 set aside every month, and after a quarter, €2,400 for one burst of addressable TV or screens on that area, once, all in. One dashboard: cost per qualified lead, per area, per week.
What I would not do under €10,000
- Five channels. Every channel added below this threshold takes from the master channel what it needs to learn.
- Changing the plan every three weeks. Every significant edit restarts learning. A small budget needs time more than a large one does.
- National TV and sponsorship. Not because they are bad, but because the first useful euro sits above your ceiling.
- Three vendors. A Meta agency, a Google freelancer, a creative studio: nobody sees the whole plan, and the whole plan is the only thing that matters at this level. We covered it in the scale-up gap: at seed stage, you buy few things, but you buy them well.
Measure simply, decide fast
No marketing mix modelling at €8,000 a month, no multi-touch attribution. One number: cost per qualified lead, or per first order, per area. And one test everyone understands: the activated area against a twin area left alone. If the difference is not visible to the naked eye in your sales, media has not reached critical mass yet, and the answer is not to add a channel, it is to tighten the area.
What I tell founders who have "a small budget"
You do not have a small budget. You have a budget that demands choices. One master channel sized to learn, one net channel, one area where you carry weight, three creatives that prove, and offline in bursts when the pot is full. With that, €8,000 a month does more than a €30,000 account scattered across six lines. I have seen it often enough to write it down.
If you want to see what your budget does under this logic, run it through our media mix simulator, then bring the result to our free audit: together we look at how many lines you fund today, and how many could actually learn.
